Chapter 30

Prostate : an Overview 30 Is Sentinel Node Detection Helpful in the Curative Treatment of Cancer?

turator lymph node dissection is currently generally General Remarks referred to as the ªgold standardº when positive nodes are suspected, followed by prostatectomy in accounts for 21% of all the case of confirmed prostate cancer. in the male population in Germany (Schçssler et As a rule of thumb, it is generally accepted and al. 1993) and has similar rates of incidence in well documented that cancer-positive aortic nodes most countries in the Western world. are connected with positive pelvic nodes and that This cancer type has long been mentally con- at least in these cases it must be accepted that the nected with ªold men's diseases,º and the level of condition is incurable. commitment to research aimed at the development As much as 20 years ago three different options of new diagnostic and therapeutic approaches to it for prostate cancer treatment in stage D1 were tested has therefore been relatively low. Now, however, (radical prostatectomy, extended radiotherapy, and with increasing survival rates even in older age hormone therapy). The median survival in all groups, interest in improving its diagnosis and groups was 39.5 months. None of the three treat- treatment is growing. With new trends in N-stag- ment strategies was superior in prolonging life (Kra- ing (see Wawroschek et al. 1999, 2000) in mind, it mer et al. 1981), and there was no convincing break- will be much more difficult to developconvincing through improving on this situation up to the 1990s. strategies for detection of (s) New approaches to antiandrogen- and radioche- (SLN), and it is also more difficult to dissect them, motherapy see Chapter 33. whether in isolation or together with secondary nodes within the pelvis, in the course of prostate cancer treatment than in the procedures used in breast cancer or malignant melanoma treatment. Serum Values of Prostate-specific Antigen The difficulties are connected mainly with prob- and Prostate Acid Phosphatase lems in administering contrast agents for sentinel as Indicators for Cancer, Metastatic Spread node detection and orientation in the local topog- and Cancer Recurrence raphy of the lymphatic network structures of the pelvis. Bluestein et al. (1994) follow from their investiga- In addition, the problem is rendered more diffi- tions on 1632 patients that prostate-specific anti- cult because not only the lymph nodes must be gen (PSA) is the best predictor of pelvic lymph dissected, but also the network of lymphatics in node metastases (P<0.0001). continuity with the prostate gland, which can con- tain cancer cells or clusters. Invasive and noninvasive imaging techniques have been flawed by unacceptably high false-posi- Definition of the Degrees of Malignancy tive and false-negative rates in most approaches in Gleason's (Scoring) System (Loening et al. 1977; Wilson et al. 1977; McCarthy and Pollak 1991; Schçssler et al. 1993). This section starts with a detailed analysis of the In view to this fact and since we have no uniform grading strategy (characteristic features of the sub- and highly developed concept for detection of senti- groups). nel node(s) that is actually practiced by the majority The predictive power is enhanced by consider- of our urologists at present, systematic iliac and ob- ing the Gleason grading (scoring) (P<0.001) and 450 Chapter 30 Prostate Cancer: an Overview

Table 1. Gleason grading system for prostatic adenocarcinoma: histologic patterns

Pattern Peripheral Stromal Appearance of Size of glands Architecture Cytoplasm borders glands of glands 1* Circumscribed Minimal Simple, round, Medium, Closely packed Similar to that expansive monotonously regular in benign growth replicated epithelium 2* Less circum- Mild, with def- Simple, round, Medium, less Loosely packed Similar to that scribed; early inite separa- some variabil- regular rounded in benign infiltration tion of glands ity in shape masses epithelium by stroma 3A Infiltration Marked Angular, with Medium to Variable More basophil- variation in large packed irregu- ic than in pat- shape lar masses terns 1 and 2 3B Infiltration Marked Angular, with Small Variable More basophil- variation in packed irregu- ic than in pat- shape lar masses terns 1 and 2 3C Smooth, Marked Papillary and Irregular Round to elon- More basophil- rounded cribriform gated masses ic than in pat- terns 1 and 2 4A Ragged Marked Microacinar, Irregular Fused, with Dark infiltration papillary, and chains and cribriform cords 4B Ragged Marked Microacinar, Irregular Fused, with Clear (hyper- infiltration papillary, and chains and nephroid) cribriform cords 5A Smooth, Marked Comedocarci- Irregular Round to elon- Variable rounded noma gated masses 5B Ragged Marked Difficulty in Irregular Fused sheets Variable infiltration identifying and masses glands' lumens * In foci suspicious for cancer (pattern 1 and 2) positive reaction using the antibody P5O4S immunostaining helps to certify cancer diagnosis

the local clinical stage of the primary together Recently the WHO published the fascicle ªTumors (P<0.001). The basic patterns leading to the Glea- of the Urinary System and Male Genital Organsº. son scores are described in Table 1. The criteria of Gleason patterns and scoring of The main criteria for delineation of the sub- prostate are summarized in the following grades are comprehensively summarized in Fig. 1 overview. The criteria of the main patterns are according to the original publications about Glea- compatible with the original description, presented son's scoring system. The subentities are still more in Table 1. precisely characterized by own drawings (Fig. 2).

WHO-Classification 2004

Gleason pattern 1 Well circumscribed of closely packed glands, no infiltration 1+1 =2 =Gleason score (local in situ condition).

Gleason pattern 2 Round or oval glands, loosely arranged, not as uniform as in pattern 1. 2+2 =4 =Gleason score Facultative minimal invasion, glands of intermediate size, with less vari- ation than in Gleason pattern 3. Correlations with Molecular Biological and Clinical Parameters 451

Gleason pattern 3 Most common pattern: Higher degree of infiltration distance of 3+3 =6 =Gleason score glands more variable, more often malignant glands between preexis- tent. Often angular and small glands besides large ones. Cribriform pattern is rare and difficult to distinguish from cribriform high grade PIN.

Gleason pattern 4 Glands fused, cribriform, may be poorly defined, partly not separated 4+4 =8 =Gleason score by stroma, edges of fused glands scalopped, thin strands of connec- tive tissue, loss of lumina in cribriform parts. Hypernephromatoid pattern, a rare variant.

Gleason pattern 5 Almost complete loss of glandular lumina, growth in solid sheets and 5+5 =10=Gleason score strands or as single cells. Comedonecroses may be present.

If the tumor has only one pattern, Gleason score is obtained by doubling the pattern number (see scheme). If there exist two different pattern, the two pattern numbers are summed up, e.g. 3 plus 4=7=Gleason score. If there is found a tertiary pattern, prognosis worsens additionally.

Correlations with Molecular Biological and Clinical Parameters

Investigations carried out by Conrad et al. (1998) in 345 cases showed that the amount and distribution of undifferentiated Gleason grades (pattern 4 and 5) cancer in the were the best predictors of the lymphatic spread, followed by serum PSA. In classification and regression analyses, nearly 80% of the patients who had Gleason score 4 or 5 disease in 3 or fewer biopsies and did not have a predominance of high-grade cancer in any can be classified as low-risk patients. Positive nodes were found in this groupin only 2.2% (95% confidence interval, 0.8±4.7%). The authors con- clude that the sextant biopsy principle enhances the predictive accuracy of algorithms, which define the probability of lymphatic spread. However, use of PSA and prostate acid phospha- tase (PAP) values does have some limitations: · Oesterling et al. (1988) concluded from their studies that preoperative levels of PSA are not sufficiently reliable for the prediction of the Fig. 1. Prostate cancer scoring according to Gleason. See Ta- ble 1 for details of the patterns pathological stage in patients with early cancer. But they confirmed the findings of other labora- tories, demonstrating that PSA is a sensitive tu- In order to give an impression of the histo- mor marker for the detection of residual disease pathological differences of high, moderate, and after radical prostatectomy and subsequent re- low differentiation cancers, Figs. 3±7 show the currence of the cancer in long-term follow-up characteristic histopathological features of each. studies. 452 Chapter 30 Prostate Cancer: an Overview

Fig. 2a,b. Analysis of the Gleason pat- terns and cytology of prostate cancer based on growth pattern and cytological criteria derived from the top-view draw- ing presented in original publications

1 Circumscribed, highly differentiated

Less circumscribed, early invasive

Glands medium to large, partly angular

Small, partly angular and ramified glands

Papillary and cribriform pattern, irregular, round to elongated cancer cell formations, cytoplasm more basophil than in patterns 1 and 2

a

· Primarily in high-grade cases, or in the course These facts must be given due consideration by of dedifferentiation, the PSA values can be low urologists, both during primary treatment and or decrease in spite of persisting or recurrent also in the follow-upof their patients, but with a cancer (Partin et al. 1990). view to the SLN concept and the performance of the necessary clinical studies the results are of spe- Conversely, unpredictable contributions from pros- cial interest. tate parts with benign prostate compo- It was found that PSA and PAP correspond to nents or foci with inflammation leading to in- histopathologically detected metastatic lymph node creased release of PSA can moderate the PSA level, involvement. causing increased serum values. Correlations with Molecular Biological and Clinical Parameters 453

Fig. 2b

Microacinar, papillary, cribriform patterns, irregular with chains and cords, cytoplasm dark

Microacinar, papillary, cribriform patterns, fused with chains, cytoplasm clear, hypernephroid

Comedocarcinoma, round to elongated cancer cell formations, cytoplasm variable

Soliud fused sheets, difficulty to identify glands

b

In the course of our daily routine work the P-val- Table 2. Relationship of levels of prostate-specific antigen ues of correlations of PSA and PAP in uni- and mul- (PSA) and prostate acid phosphatase (PAP) in cancer tissue tivariate analysis are of special interest (Table 2). values to histopathological grade, DNA ploidy and T-catego- ry. The following values were found In multivariate analysis the serum PSA level was the most powerful independent prognosticator fol- N category to serum PSA level P<0.001 lowed by the Tcategory, tissue PAP and tissue PSA. to histological grade P<0.001 But there are limitations to the detection of to tissue PSA P<0.001 early cancer development, as already mentioned above. to tissue PAP P<0.004

to Tcategory P<0.005

to DNA ploidy P<0.002 454 Chapter 30 Prostate Cancer: an Overview

Figs. 3±7. Degrees of malignancy of pros- tate cancer: main types Fig. 3. Highly differentiated prostate can- cer corresponding to Gleason pattern 2, corresponding to Gleason score 2+2=4. Note the fairly isomorphic glandular structures

Fig. 4. Well-differentiated prostate cancer with fairly isomorphic cancer cell nu- clei, corresponding to Gleason score 3+3=6 low rate of mitoses, only focal branching of the glands

Fig. 5. Moderately differentiated prostate cancer with more pronounced variation of the glands. Gleason pattern 4, corre- sponding to Gleason score 4+4=8 Relation Between and Cancer 455

Fig. 6. Anaplastic prostate cancer with high grade of polymorphism and high rates of mitosis. Gleason score 10

Fig. 7. Anaplastic prostate cancer with high grade of tumor cell dissociation and nerve sheath invasion. Gleason score 10

are probably unlikely to reach clinical significance Importance of Tumor Volume because the doubling time of this cancer is too to Clinical Significance long. in Treatment of Prostate Cancer

Minimal Lesions Relation Between Dysplasia and Cancer

Stamey et al. (1993) tried to calculate the probabil- In addition, McNeal (1993), also from Stanford ity of having a diagnosis of prostate cancer within University, investigated with dysplastic a man's life. The prostates of 139 consecutive blad- changes of glands in the biopsy material. In 48% der cancer patients who had undergone cystopros- of these he found foci fulfilling the criteria for tatectomy were examined. Prostate cancer was prostate cancer, and in 3 of the 107 cases investi- found in 55 patients (40%), larger cancers being gated grade 4 cancer (low degree of differentiation) detected only in 8%. The cancers ranged in volume was found. Among patients in whom microcarci- from 0.5 to 6.1 ml. These results allowed the con- nomas were diagnosed, dysplasia was found in clusion that prostate cancers smaller than 0.5 ml other parts in 81%. 456 Chapter 30 Prostate Cancer: an Overview

These data may also help in decision making Gleason Score (Grading) in Ultrasound-guided when a catalogue of the indications for sentinel Biopsies Related to Results in Prostatectomy node evaluation is developed. Specimens Carlson et al. (1998) also examined the correla- tions between Gleason scores in biopsy specimens In their comparative studies of ultrasound-guided and the diagnoses made in prostatectomy speci- biopsies and prostatectomy specimens conducted mens. They investigated 106 cases and obtained in 289 cases, Gregori et al. (2001) obtained the fol- the results summarized in Table 5. The positive lowing results summarized in Tables 3 and 4. predictive values of the different Gleason scores It is very interesting that with whole-prostate are displayed in Table 6. evaluation (specimen investigations) as reference, upstaging is necessary in about 40%, that among cases with unilateral positive biopsies bilateral can- Division of Grade II (WHO) Cancers into Favorable and cer infiltration is found in about 65%, and that Unfavorable Subgroups Supported by Gleason Grading among cases with bilateral positive biopsies intra- capsular cancer infiltration is found in about 66%. Lilleby et al. (2001) tried to differentiate more pre- cisely between favorable and unfavorable subtypes of grade II cancers graded by means of the WHO criteria compared with the Gleason-graded sub- Table 3. Predictive value of Gleason score, up- and down- grading (according to Gregori et al. 2001) groups. The review of specimens from 178 patients Recorded Identical Up-grading vs yielded the numbers of cases in the different no. of patients Gleason score down-grading groups shown in Table 7. in the biopsies related to histo- logically investi- Separating patients with a Gleason score of 7 gated specimens (score 3+4 vs 4+3) led to a two-tiered Gleason grouping (88 patients in the favorable group and 289 126 (43%) 118 (40.8%) vs 43 (14.8%) 90 in the unfavorable group). The authors concluded that equal allocation of patients to subgroups based on the Gleason system helps the clinician to overcome the dilemma of Table 4. Specimen-related histopathological results in cases overrepresentation of grade II patients, which does with uni- and bilateral positive biopsies (according to Gre- gori et al. 2001) occur with the WHO grading. Subgrouping grade II cancers by means of the Unilateral Bilateral Organ Capsule Gleason grading system appears to make separa- positive disease confined penetration biopsy vs tion more of a possibility, for instance with a view cancer on to an additional sentinel node search and referral one side to radiotherapy. 193 127 142 41 Altay et al. (2001) found that high Gleason (66.7%) (65.8%) (73.5%) (26.4%) scores together with elevated PSA levels (>10 ng/ ml) involved a high risk of extraprostatic cancer 46 (33.1) extension, in some cases with seminal vesicle in- volvement (P<0.05). Bilateral Intra- Overall Postero- Apical positive capsular positive lateral capsular biopsy cancer margins capsular penetra- (n) vs speci- penetra- tion Grading Errors in Gleason Grading Evaluations men tion confined Compared with a Database

96 64 14% 52/83 28/83 King and Long (2000) investigated the question of (66.6%) (61.4%) (33.7%) grading errors. A pooled database from ten series 32 (n=2687) served as a baseline for comparison (33.3%) aimed at checking the accuracy of Gleason score Simultaneous Lymphogenous and Hematogenous Metastatic Spread? 457

Table 5. Accuracy of Gleason biopsy grading

No. of Correlation of biopsy Correlation All patients Under-grading Over grading patients with radical within one grade correlated within prostatectomy 2grades 106 72 (68%) 103 (97%) 106 (100%) 26 (25%) 8 (8%)

Table 6. Predictive values

Percentage of positive predictive values Well Moderately Moderately Poorly differentiated differentiated to poorly differentiated differentiated Gleason 5 6 7 2±4 5±6 7 8±10 66% 67% 71% 70%

Table 7. Distribution of prostate cancer cases (n=178) within the Gleason grading scheme and rate of grade 2 cancers in this collective (WHO grading)

Total no. of cases Gleason <7 (n) Gleason 7 (n) Gleason 8±10 (n) Grade II WHO (n) 178 44 58 76 130

Table 8. Deviations of grading prostate cancer cases using · Scores of 5 and 6 in 47% the Gleason scale from the database · Score of 7 in 47% Within 1 point in 93% P=0.029 · Scores of 8±10 in 25% Compared with database in 78% P=0.029 Underestimation of the growth pattern was also Upgrading of biopsies in 35% P=0.19 observed: · Pattern 2 in 32% Compared with database in 43% P=0.19 · Pattern 3 in 39% Gleason 7 exact match in 78% P=0.07 · Pattern 5 in 30% Gleason 7 database in 20% P=0.07 The authors recommend training programs and courses to improve the quality in grading-practice and to reduce interobserver differences. grading. With the biopsy technique used an exact Gleason score match was achieved in 57% of the cases, a mean of 42% (P=0.055) compared with Simultaneous Lymphogenous the pooled database (PD). Table 8 shows the P-val- and Hematogenous Metastatic Spread? ues for deviation from the database. King and Long's conclusion was that: ªSampling effects may The fact that autopsy studies both in the United contribute significantly to grading errors in pros- States and in Japan show identical rates for meta- tate needle biopsies.º static involvement of lymph nodes and of bone Koksal et al. (2000) found that Gleason grading marrow (see Fig. 11) can be interpreted as showing error compared with the grading of prostatectomy that regional lymphatic spread and hematogenous specimen grading was highest in highly differen- retrograde transport to the column and other bone tiated cancers. This can be easily understood on regions are closely connected. the basis of primary multifocality clonal selections. The additional finding of lung metastases in Allsbrook et al. (2001) evaluated the interob- two-thirds of cases with lymph node and bone server reproducibility of the Gleason grading: con- marrow involvement can be explained by the com- stant undergrading was observed in the Gleason munication of the venous periprostatic and prever- scores, at the following rates: tebral venous plexus with the azygos veins, which 458 Chapter 30 Prostate Cancer: an Overview

are directly connected with the vena cava inferior. Smith et al.'s (1983) investigations already make It follows from this that it is only in the very early it clear that locoregional metastatic spread de- stages of local lymphatic spread that there can be pends heavily on how malignant the primary is any hope of local curative lymphatic operative tu- (Tables 9, 10). These data help in preoperative cal- mor clearance. culation of the risk of locoregional cancer progres- These reflections are important because they sion. can influence the development of a catalog of indi- In addition, the following data published by cations for sentinel and pelvic lymph node extirpa- Walsh et al. (1994) can be used to obtain a pro- tion and investigation projects. That means that it spective assessment of 10-year survival, depending is very important to perform bone and lung-inves- on Gleason grade and capsule infiltration or per- tigations before surgical procedures are started. foration by the cancer (Fig. 8). These data allow easier formulation of indica- tions for a differentiated sentinel lymph node Significance of Degree of Malignancy (SLN) search based, for instance, on the new re- sults and developments published by the Augsburg and Number of Biopsies Taken for N-staging research group (Wawroschek et al. 1999, 2000). and the Sentinel Lymph Node Concept Danella et al. (1993) reported that in their group of patients with clinically localized prostatic can- According to the experience of pathologists, foci of cer, 5.7% had nodal metastases. The predictive val- high-grade prostate cancer can be found in a set of ue in patients with prostate-specific antigen levels the typical six-point biopsy series. This leads to elevated to >40 ng was 53%. The authors empha- the conclusion that high-grade cancers develop as size that laparoscopic lymph node dissection subclones in the course of cancer extension. should be the method of choice in view of the low McNeal (1993) investigated this problem again rate of nodal metastases. and demonstrated the derivation of high-grade In the context of these important facts, in addi- cancers as more or less extended subclones in low- tion to local regional operative pelvic cancer clear- grade cancers within the prostate gland. ance, if the near future were to see improvements Usually the majority of nonextended microcar- in prostate cancer grading including the emer- cinomas develop from preexistent dysplasia. Inves- gence of important prognostic factors indicative of tigations of nonextended early cancer of the pros- likely proliferation and blood vessel invasion with tate show areas with high-grade cancer in only 3% a high predictive value these would be greatly ap- of cases. preciated. Knowledge of these proportions seems to be im- portant for the development of a sentinel lymph node concept that is suitable for application in the treatment of prostate cancer, especially with the aim of finding the limits for N-staging in early cancer cases with a low grade of malignancy.

Table 9. Distribution of tumor grade by stage

Stage Well differentiated Moderately differentiated Poorly differentiated Total no. of patients

n (%) n (%) n (%)

A1 28 (68) 12 (29) 1 (3) 41 A2 7 (21) 19 (58) 7 (21) 33 B1 53 (34) 94 (60) 9 (6) 156 B2 27 (18) 106 (69) 21 (13) 154 C 10 (15) 44 (64) 14 (21) 68 125 (28) 275 (61) 52 (11) 452 Basic Research for Complete Pelvic Lymph Node (N-) Staging 459

Table 10. Incidence of pelvic node by histological grade and clinical stage [from Smith et al. (1983)]

Stage Well differentiated Moderately differentiated Poorly differentiated Total no. of patients

No. (%) No. (%) No. (%) No. (%)

A1 0/28 0/12 0/1 0/41 A2 0/7 5/19 (26) 3/7 (43) 8/33 (24) B1 2/53 (4) 13/94 (14) 3/9 (33) 18/156 (12) B2 5/27 (18) 29/106 (27) 9/21 (43) 43/154 (28) C 5/10 (50) 18/44 (41) 13/14 (93) 36/68 (53) 125 (28) 275 (61) 52 (11) 452 (23)

Fig. 8. Actuarial probability of freedom from recurrence of elevated PSA in rela- tion to stage and Gleason grade. These data are based on follow-up studies of 955 men treated at Johns Hopkins Hos- pital by Walsh et al. (1994)

Is the Primary Cancer Detection Rate Higher Basic Research for Complete with Twelve Biopsies than with Six? Pelvic Lymph Node (N-) Staging

Prospective randomized studies carried out in 244 Metastatic Involvement of the Main Node Groups men by Naughton et al. (2000) and comparing the and Statistical Evaluations of Side Differences impact of the number (6 versus 12) prostate biopsies on cancer detection did not show higher detection Weingårtner et al. (1996) tried to find mean values rates when 12 core biopsy specimens were taken in an attempt to answer the question of how many and investigated in a screening program (P=0.9). lymph nodes need to be investigated to give the At around the same time a report of a study complete metastatic profile in the pelvis. They per- conducted by Ravery et al. (2000) to evaluate 303 formed investigations both on cadavers and in liv- consecutive cases showed an overall detection rate ing cancer patients. The means found were 22.7 for prostate cancer of 38.9% in their patient popu- for the cadavers (Ô10.2, range 8±56) and 20.5 for lation, the rate being higher by 6.6% when more the prostate cancer patients (Ô6.6, range 10±37). biopsy investigations were done (6.5% in men with In the patient group, involvement of pelvic PSA 10 ng/ml or less and 7% in men with PSA lymph nodes was significantly more frequent on >10 ng/ml). (For more information on increases in the left side (see Fig. 9). Lymph nodes were more detection rates see also: Eskew et al. 1996; Beurton frequently enlarged in cancer patients than in con- et al. 1997; Ravery et al. 1999.) trols, regardless of whether they were cancer infil- 460 Chapter 30 Prostate Cancer: an Overview

Survival Rate Versus Lymph Node Metastasis

In a comment on Gervasi's results (see Table 11), Paulson (1989) suggested that hematogenous met- astatic spread develops more or less in parallel at the same time. This suggestion seems to hold true for many cancer cases progressing by the hematog- enous route, but we cannot at present relate this assumption to particular cancer cases on the basis of specific clinical and molecular biological fea- tures (Fig. 10). However, in spite of these tentative conclusions, locoregional pelvic cancer clearance (tumor-free margins of the primary, complete node dissection) should be consistently developed in an indepen- dent way. Before a sentinel node search is conducted dis- tant metastases should be excluded. The frequen- cies and sites of hematogenous metastases are listed in Fig. 11. It is worth emphasizing that in the prostate the lymphatics follow the larger lymph channels leav- ing the prostate posteriorly and then spread pri- Fig. 9. Main pelvic lymph node groups that can be involved in metastatic spread. Note that the left-sided lymph nodes marily to the perivesical, hypogastric, obturator, are more frequently involved presacral and presciatic and obturator lymph nodes as the first stations (McLaughlin et al. 1976). These results are already quite old and are trated or tumor free. Therefore, enlarged lymph not fully included in the current routine principles nodes cannot be used for diagnostic purposes as of lymph node clearance. they are in breast or lung cancer, e.g. in imaging This posterior extracapsular region could be one investigations. These results indicate that it is im- of the points where the contrast solution for SLN la- possible to use CT or MRI to diagnose metastatic beling could be injected. Systemic iron oxide (Siner- cancer involvement of the pelvic nodes. em) injection should also be discussed again as a The rates of metastatic lymph node involvement possibility for labeling the first (sentinel) stations. from prostate cancer in the different pelvic posi- In Schçssler's investigations (1993) in cases with tions (iliac and obturator) are shown in Fig. 9. palpable disease (stages B1, B2, C), 36% had lymph In Gervasi et al.'s (1989) series of 511 prostate node involvement. This value corresponds to many cancer patients, N0 and N+ patients were com- other series investigated in different clinics (Flocks pared for metastatic lymph node involvement over et al. 1959; McCullough et al. 1974, 1977; McLaugh- a median follow-upof 8.6 years; the P-value was lin et al. 1976; Bruce et al. 1977; Wilson et al. 1977; less than <0.00005 (Table 11). Freiha et al. 1979; Brendler et al. 1980; Grossman et The hematogenous metastatic rates determined al. 1980; Fowler and Whitmore 1981; Catalona and for N+ patients were: Stein 1982; Donohue et al. 1982). Widely different frequencies have been recorded N1 80% for metastases to the external iliac lymph nodes: N2 84% N3 88% Nicholson and Richie (1997) 17% McLaughlin et al. (1976) 10% Arduino and Glucksman (1962) 57% Bruce et al. (1977) 54% Schçssler et al. (1993) 30% Basic Research for Complete Pelvic Lymph Node (N-) Staging 461

Table 11. Actuarial all-cause and cancer-specific survival rates at 5, 10 and 15 years by extent of nodal metastases (percent Ô2 SE) (from Gervasi et al. 1989)

All-cause survival rate Cancer-specific survival rate

5 Years % 10 Years % 15 Years % 5 Years % 10 Years % 15 Years %

N0 90Ô3 62Ô6 36Ô14 98Ô2 83Ô6 70Ô13 N+ 65Ô8 32Ô9 6Ô10 75Ô7 43Ô11 21Ô17 N1 68Ô15 40Ô19 ± 82Ô13 60Ô19 ± N2 63Ô11 22Ô11 ± 72Ô10 34Ô15 ± N3 69Ô17 35Ô22 ± 75Ô17 42Ô24 ±

Fig. 10. Distant metastases (actuarial rate) by presence and opment of distant metastases was parallel in each of the extent of nodal metastases. Vertical bars indicate 95% confi- three subgroups of N+ cancer patients (see Gervasi et al. dence intervals (meanÔSE). Although distant metastases 1989, Table 11) appeared less rapidly in the N1 subgroup, the rate of devel-

Fig. 11. Sites and frequencies of metasta- ses from prostate cancer and most com- mon sites of prostate cancer metastases found at autopsy in the United States and Japan [from Bostwick and Eble (1993) with permission] 462 Chapter 30 Prostate Cancer: an Overview

These wide differences (10% to nearly 60%) in 1. Evaluation of grade of malignancy with rating the rates of external iliac lymph node metastases of nuclear grade plus, if possible, evaluation of are disturbing when we are trying to collect clear- the percentages of cancer cells in DNA synthesis cut information for use as a basis for the develop- and in the proliferative compartment, using the ment of a SLN concept. antibody MiBI and the CASS 2000 machine for The percentages of cases upstaged to D1 from the determination of percentages. the different lower stages according to Schçssler 2. Proof of cancer infiltration and perforation of (1993) evaluations of lymph node involvement are the prostate capsule; when these pathohistologi- as follows: cal, biologically based parameters allow the con- clusion that a high-grade or high-risk cancer is present SLN staging should be performed. A2 27% · In addition to the standard dissection involving B1 15% the nodes along the common iliac artery, the B2 52% external iliac artery, genitofemoral nerve, hypo- C 50% gastric vessels and obturator fossa also need in- vestigation (Flocks et al. 1959; Arduino and The percentages for stages B2 and C are very Glucksman 1962; Barzell et al. 1977; Bruce et al. high. Better information on prognostic factors is 1977; Freiha and Salzman 1977; Nicholson and thus long overdue. Richie 1977; Grossman et al. 1980; Fowler and When all aspects are considered together, two Whitmore 1981; Donohue et al. 1982). principles seem to be the most important indica- · An extended dissection also includes the nodes tors of regional lymph node involvement and thus in the presacral and lateral sacral areas (see also to constitute indications for a search for sentinel Wawroschek's, Vogt's and Harzmann's First nodes and in addition, when the sentinels are pos- Approach to Detection of Sentinel Nodes in itive, for pelvic lymphadenectomy. These two most Prostate Cancer in this chapter). It is note- important principles are: worthy that some investigations have indicated

Fig. 12. Inverted-V peritoneotomy ac- cording to See et al. (1993). The perito- neal flapis dissected superficially and folded back along the dotted lines shown in the schematic Intraoperative and Postoperative Lymph Node Staging in the Treatment of Prostate Cancer 463

that in approximately 13±15% of cases these ary prostatectomy. Nonetheless, the application of locations are the only ones involved in the met- this method has limitations. astatic process. · Trocars (three 11 mm and two 5 mm in diame- · The limited dissection procedure is derived ter) need to be inserted at five locations , dis- from the standard method. The lateral border is tributed over the lower parts of the abdomen. limited by the lateral margin of the lateromedial · This method has a similar complication rate to aspect of the external iliac artery instead of the open biopsy. genitofemoral nerve (Brendler et al. 1980; Paul- · The fundamental problem of cancer progression son 1980; Lieskovsky and Skinner 1983; Gervasi after nodectomy in cancer-positive cases is not et al. 1989; McDowell et al. 1990). The common improved. iliac artery is left undissected, which is easily Therefore, improved local clearance in the sense of understood in view of the difficulty of imple- the SLN concept has not yet been reached. menting lymph node and lymph vessel clearing principles in the pelvic region. · A fourth possibility is obturator fossa node dis- section (Kozlowski and Grayhack 1987; Winfield Intraoperative and Postoperative Lymph Node and Kavoussi 1991; Winfield et al. 1991), which Staging in the Treatment of Prostate Cancer is a simpler procedure with low morbidity rates. Intraoperative lymph node staging in frozen sec- · In comparison with linear peritoneotomy, an in- tions needs close cooperation between pathologists verted-V peritoneotomy, which is designed to and their technicians, because sometimes microfo- allow wider exposure of the target tissue, im- ci of cancer infiltration are visible even at the sec- proves the nodal yield of laparoscopic lympha- tion surface. Double investigations of frozen sec- denectomy (Fig. 12). tions from each lymph node slice and of imprint cytology of each slice of the nodes can increase ef- It may be that inverted-V peritoneotomy will be ficiency and lower the false-negative rate. one of the keys to complete lymphadenectomy in However, we should always keepin mind that the the further development of a sentinel node concept quantity of tissue lost is much greater in frozen sec- for treatment of prostate cancer. It is clear that tions than in paraffin embedding and paraffin sec- high complication rates of standard pelvic lymph tioning. Therefore, when two-stepsurgery is node dissection with complication rates nearly planned (node evaluation and prostatectomy) basing equal to those of laparoscopic lymphadenectomy a decision for prostatectomy exclusively on frozen (30±32% for both methods) complicate the devel- sections of the lymph nodes is of restricted value. opment of sophisticated methods for SLN dissec- This question becomes even more difficult when tion programs. The complications are: hemor- attempts at development of a SLN concept are rhage, bowel laceration, ureteral injury, and devel- afoot, because sentinel nodes, as the key nodes in opment of lymphocele. metastasis, must be investigated in serial sections, a technical procedure that is more reliable and eas- ier to perform when the paraffin technique is used. In recent times, results of some screening- and comparative studies have been published to allow Is Laparoscopic Lymph Node Staging Equivalent evaluation of the efficiency of different methods. to Open Pelvic Lymph Node Dissection? Epstein et al. (1986) presented their N-staging data based on a series of 299 cases. They empha- Compatibility with a Pelvic SLN Concept sized that the detection of metastases by the frozen section technique is quite precise. The average ex- Parra et al. (1992) state that laparoscopic node ex- tension of the metastases was 2.5 mm, and the cision gives node detection rates identical with false-negative rate was 3.5%. those obtained with open operative dissection. No In a frozen section lymph node analysis carried more nodes were found after laparoscopic node out by Davies (1995) a 90% diagnostic efficiency dissection from the surgical margins by open re- could be reached, which is similar to that of MRI operation in patients who had undergone second- analysis (88%). However, the aggregated sensitivity 464 Chapter 30 Prostate Cancer: an Overview

of MRI was less than 67% that obtained with fro- Table 12. Rates of lymph node involvement in different zen section analysis. stages of prostate cancer in the collective published by In our opinion, the histopathological efficiency Schçssler et al. (1993) can still be increased by intraoperative serial sec- Stage No. of Cases with positive nodes tion analysis using an immunohistochemical tech- cases No. % nique newly developed by Hæfler and Nåhrig, in which the staining procedure takes only 12±20 min Ul a 28 0 0 (see also Chapter 17). A2 11 3 27 As in staging procedures used for other cancers [see Chapter on Melanoma (Chapter 25)], Okegawa B1 21 3 15 et al. (2000) also performed RT-PCR studies of the B2 21 11 52 dissected lymph nodes for staging prostate cancer C6350 by preliminary investigations, to detect early lymph node metastases, and to define the indica- a Adenocarcinoma of prostate detected by ultrasound only tions for prostatectomy more precisely. Positive re- and in addition with view to the literature [see also Ta- ble 13, also from Schçssler et al. (1993)] sults were obtained in 11% of stage pT2a and in 25% of stage pT2b cases. Invasion of prostate cancer into seminal vesicles did not significantly alter the frequency of involved nodes (61% versus 70%), as published by Barzell mor progression than B1 cancers with involvement et al. (1977). of less than half of a lobe of the prostate. The reason why the Whitmore scaling cannot be used as a basis of decision about in which cases lymph node staging will be helpful for further de- cision-making (sentinel node labeling, prostate Is Whitmore's Staging Scheme Established adenectomy, etc.) lies in values given (for example) in 1984 Still Compatible by Schçssler et al. (1993) (Table 12). with Our Current Knowledge? A decision in favor of lymph node dissection is made possible by the fact that lymph node in- The establishment of indications for a search for volvement is more frequent (according to collective SLNs is at least partly hampered by the fact that evaluations) in A2 than in B1 stages (26% vs the clinical staging devised by Whitmore in 1984 16%). However, the background to this marked is incompatible with tumor biology in important discrepancy is that all extensions and all grading points. To help readers understand the critical groups are included under A2! This seems a poor points, first of all the topographically oriented basis for preoperative selection of the cases in scheme of Whitmore should be described: which lymph node involvement must be assumed. The percentages of different rates of lymph A2: Any quantity of cancer within the prostate, node involvement published by different working without discrimination between the different groups on the basis of the Whitmore staging (A2± degrees of malignancy, meaning by implication C), are listed in Table 13. Some investigations give high-grade cancer types. the impression that lymph node involvement is B1: Cancer nodule involving less than half of a found almost exclusively in cases in which the can- lobe. cer has broken out through the prostate capsule. B2: Cancer nodule involving more than half of a However, a systematic search for nodal microme- lobe or bilateral cancer nodes. tastases by means of modern immunohistochem- C: Local extension of any cancer beyond the pros- ical techniques reveals micrometastases in 16% of tate gland. pT3N0 cases and also in lower pT stages (Gomella D1: Prostate cancer with positive pelvic nodes. et al. 1993; Moul et al. 1994). D2: Prostate cancer with distant metastasis. In the context of these findings we must see that many case series are incorrectly staged. This It is clear that inclusion of all high-grade cancers knowledge has a marked influence on our new in stage A2 must give worse results in terms of tu- approaches to developing a sentinel node concept. Is Whitmore's Staging Scheme Established in 1984 Still Compatible with Our Current Knowledge? 465

Table 13. Laparoscopic standard of pelvic node dissection for prostate cancer. Incidence of positive nodes according to clin- ical stage: review of the literature according to Schçssler et al. (1993)

Reference Clinical stage

A2B1 B2 C

Total Patients Total Patients Total Patients Total Patients no. of with no. of with no. of with no. of with patients positive patients positive patients positive patients positive nodes nodes nodes nodes n (%) n (%) n (%) n (%)

Donohue et al. 59 13 (22) 179 26 (15) 99 46 (46) ± ± (1982) Freiha et al. 2 0 (0) 13 2 (15) 44 10 (22) 41 24 (58) (1977) Bruce et al. (1977) 3 0 (0) 6 0 (0) 13 5 (38) 8 6 (75) Nicholson and ± ± 26 2 (8) 14 2 (14) 6 2 (33) Richie (1977) McLaughlin et al. ± ± 19 4 (21) 17 5 (29) 24 12 (50) (1976) Brendler et al. 22 3 (14) 58 11 (19) 27 14 (52) 17 10 (58) (1980) Catalona and 9 3 (33) 49 14 (28) 28 9 (32) ± ± Stein (1982) Grossman et al. 47 25 (53) 18 3 (17) 14 4 (28) 9 5 (55) (1980) Fowler and Whit- ± ± 75 5 (7) 129 56 (43) 96 58 (60) more (1981) Gervasi et al. 130 29 (22) 165 35 (21) 100 37 (37) 116 51 (44) (1989) Wilson et al. 8 0 (0) 36 5 (14) 29 9 (31) 19 10 (52) (1977) McCullough et al. ±±±±± ±2719(70) (1977) Flocks et al. ±±±±± ±382144(37) (1959) Totals 280 73 (26) 644 107 (16) 514 197 (38) 745 341 (46)

The metastatic rates in a series of 521 patients tastases was hardly more than half of Schçssler's investigated by Petros and Catalona (1992) were as rate. follows: With a view to tumor malignancy grade, in to- tal metastases were detected in only 10% of low- A1 0% grade cancers, in 24% of moderately differentiated A2 3.3% cancers, and in 54% of high-grade tumors (poorly B1 5.3% differentiated). B2 9.7% The fact that the rate of metastases from well- differentiated cancers (A1=0%, B1=4%) is very In a series of 452 cases investigated by Smith et al. low led to the conclusion that lymphadenectomy

(1983), especially in the B2 group the rate of me- should not be performed in these cases even at the 466 Chapter 30 Prostate Cancer: an Overview

beginning of the 1980s. Conversely, in poorly dif- time. Therefore, a wait-and-see strategy is mostly ferentiated cases in stage C, metastatic behavior adopted with the background intention of per- can be assumed in more than 90% (93% according forming palliative if necessary. to Smith et al. 1983). In these conditions there is only one choice, as In conclusion, on the basis of their own results has long been the case in many clinics: and those in the literature, Petros and Catalona · Investigation of the six routinely taken biopsies, (1992) reflect on genuine changes in our estimates with the possibility of dividing the cases into of the stage at which prostate cancer is currently two main groups diagnosed. The authors emphasize that the main ± Low-risk groupwith regional lymphnode in- factors in the lower metastatic rates seem to be: volvement in a maximum of 2% of cases: low · Higher index of suspicion grade (I) with limited intraprostatic exten- · Earlier detection sion, no capsule perforation · More aggressive intervention to establish the di- ± High-risk groupwith regional lymphnode agnosis involvement in upto approximately 30%. · Ultrasound-guided prostate biopsies · Giving the patient information, with the aim of · More widespread screening for prostate cancer. making him feel able to give informed consent to further treatments.

This means regional lymph node investigations on Development of a SLN Concept frozen sections during the operation and prosta- in Relation to Tumor Stages tectomy in patients with negative nodes. This is the most frequently practiced concept, Evaluation of Preconditions and it does not fit in directly with the concept of sentinel node labeling. Changes are needed to Recently Huland (1998) tried to discuss the value bring about better quality of the investigations and of radical prostatectomy in node-positive cases. In of decision-making process. a first statement he declared that in the course of The sentinel lymph node concept is already last 10 years, in stages T1 and T2 lymph node in- practiced in many different ways and could be volvement has decreased from 20 to less than 7%. adapted to the prostate cancer problem as follows: In a second statement he deals with the relapse · To obtain the most clear-cut information possi- rates recorded by Abbas and Scardina (1993), ble about the localization, especially when the Catalona and Smith (1994), and Walsh et al. positions of the sentinel nodes are atypical, it is (1994): nearly 100% in stages N1. He goes on to necessary, especially in cases in the high risk say that according to these studies that have ema- group, to wait for the results obtained in paraf- nated from Johns Hopkins, Houston and St. Louis fin sections of the nodes (principally investi- Hospitals and to other studies it seems doubtful gated in the same way as in breast cancer or whether ever a patient with positive node(s) could melanoma cases). be cured by radical prostatectomy and lymphade- · Decisions about prostatectomy can then be nectomy (Abbas and Scardino 1993; Catalona and made with informed consent from the patient Smith 1994; Walsh et al. 1994). and the possibility of later adjuvant radiation Huland cites the Gæteborg studies and empha- therapy can also be discussed with him. sizes that among 514 patients 319 died from pro- gression of their prostate cancer; 131 of these 319 It has long been unclear whether the survival rate of cases (41%) needed a ªtransurethralº operation, N-positive cases can be increased by more discrim- and 98 needed palliative radiation therapy to the inating and continuous surgical removal of regional urinary bladder. lymph nodes. This is partly because no results of In 53 cases (17%) urine had to be diverted into systematic comparative studies of groups treated the cranial part of the urinary bladder. In approxi- with and without node resection, each subdivided mately 20% of cases a prostatectomy could help to into low- and high-grade groups, are available. How- avoid later severe complications that could other- ever, it is easy to understand why this is so, because wise be caused by growth of the cancer, but we it must be seen that it would be unethical not to dis- have no method of identifying the groups ahead of sect the regional lymph nodes, i.e., the iliac and ob- Current Survival Rates as a Measure of Improvements in Lymph Node Staging and Clearance 467 turator lymph nodes on both sides, in the case of erated by most patients, but is not ideal in terms of larger tumors, especially when the prostatic capsule lymphatic flow from the tumor region to the SLN(s). is shown by histological investigation to be tumor In addition, because it is often not possible to infiltrated, something that can already be detected distinguish the outline of the cancer within the in the biopsies in special cases. gland, the injection would be given directly into This dilemma is further complicated by the fact cancer tissue, which seems to be dangerous, espe- that some authors see laparoscopic and open surgi- cially in view of the increased pressure that must cal procedures for lymph node extirpation as com- follow the injection of 2 ml of the labeling-solu- peting methods. With this rather confused situation, tion, which can cause blood and lymphatic vessels the approach to looking for sentinel node(s) has to open (see also discussion of this problem in been founded on a very unclear basis. Many points Chapter 7, which is focused on breast cancer label- are still open and urgently need to be cleared up. ing procedures). The following questions arise in this context of On the other hand periprostatic injection of the the approach that is striven for: contrast solution also seems to be problematic, be- 1. Can laparoscopic and open surgical pelvic lym- cause venous blood vessels of the plexus prostati- phadenectomy be recognized as methods of cus can be opened, followed by bleeding and equal value? thrombosis (see Fig. 13). 2. Is it possible to label the cancer within the prostate gland with the aim of secondary senti- Ad 3: The influence of lymphogenic micrometastases nel lymph node labeling using the methods de- on tumor progression cannot be judged precisely at scribed by Harzmann and his group, for exam- present, because the data available are in a highly de- ple (see Wawroschek et al. 1999, 2000)? gree divergent and based on different preconcep- 3. Can labeling of sentinel nodes with atypical lo- tions. Therefore, the following options must be cations be helpful in revealing micrometastases? thought through before we can find better solutions: · Is it possible to localize small cancers very accu- Ad 1: Minimally invasive surgical methods allow rately by ultrasound and inject labeling solution excision of the iliac and obturator lymph nodes on into the interstitial non-cancer-infiltrated part of both sides for pathohistological investigation of al- the prostate gland body? most the same quality as is possible with an open · Is it possible to localize smaller foci of cancer, operative procedure. for instance in cases with multifocality, and in- If this statement were absolutely correct, the ject labeling solution into the interstitial pros- approach could be to perform histopathological tate gland body? examination of the laparoscopically dissected · Is it possible to inject the contrast solution into nodes, followed in node-negative cases by transur- the vicinity of the primary outside the gland body ethral or perineal prostatectomy. However, this but in highly reduced volumes of maximally 0.2± therapeutic approach does not allow labeling of 0.5 ml in order to reduce the interstitial pressure prostate gland with 99mTc and a search for the sen- to a high degree so as to avoid opening of blood tinel regional lymph node(s) by means of the gam- vessels, which is otherwise induced by high pres- ma probe, because SLNs in atypical locations can- sure and leads to interruption of the continuity of not be dissected by laparoscopic methods. This vascular wall structures? procedure would only be possible with open op- erative methods. In addition, there is real doubt among urologists about the comparability of mini- mally invasive and open lymphadenectomy. Current Survival Rates as a Measure of Improvements in Lymph Node Staging Ad 2: As is generally well known, it is already diffi- and Clearance by the Sentinel Node Concept cult to obtain significant prostate gland material from the six different areas of the gland, because When surgical lymph node clearance, especially in the procedure for obtaining core biopsies of the pros- cases with occult or micrometastases, needs to be tate gland is very painful for the patient. On this ba- extended, as already proposed by Wawroschek, sis, we can be sure that the injection of 2 ml colloidal Vogt, and Harzmann, the procedure should be ori- solution, for instance, into the prostate gland is tol- ented on the current survival rates. 468 Chapter 30 Prostate Cancer: an Overview

Fig. 13. Venous pelvic bloodstream: Note the intensively developed venous plexus prostaticus (PP) and plexus presacralis. The PP is closely connected to the pre- vertebral and vertebral plexuses. These connections are relevant after veins have been opened and venous blood pressure has risen in connection with prostate and pelvic lymphadenectomy, as they al- low so-called retrograde transport of cancer cells into the vertebral bone mar- row. After Leibovitch et al. (1995)

Cheng et al. (1993) published the overall surviv- Table 14. Five years respectively ten years survival rates in al rates achieved with different therapeutic princi- D1-prostate cancer patients (non-randomized study) at 631 ples (see Table 14). cases by Cheng et al. (1993) (IO intraoperative, PO post- operative) In other series, e.g. that published by Hanks (1993), the results obtained with adjuvant radia- Therapy regimen Survival tion therapy have also been limited (see Table 15). 5 Years 10 Years The data presented in Tables 14 and 15 confirm that operative procedures obviously have some Prostatectomy and orchi- 91% 78% benefit but that this is limited compared with irra- dectomy (PO, n=251) diation therapy, and suggest that orchidectomy Prostatectomy alone 91% 75% prolongs survival. (PO, n=78) In these circumstances, and in view of the fact Irradiation and orchidec- 84% 54% that by 10 years after the primary diagnosis most tomy (IO, n=97) patients have reached the mean general survival Orchidectomy alone 84% 45% age, it seems that sentinel node detection with (PO, n=60) curative intent is limited, when all stages are taken together; therefore, sentinel node detection must be focused on the early stages of cancer develop- ment with histologically occult metastases and mi- Whereas in locoregional tumor clearance in crometastases, such as those detectable only by breast cancer irradiation therapy (homogeneous ir- immunohistochemical serial sectioning, to improve radiation after wide excision of the primary) is very locoregional clearance. Such efforts may improve successful in preventing local recurrence and in gen- the statistically evaluated success rate, but must be eral in improving healing rates, the same level of seen mainly under the aspect of individual thera- success obviously cannot be reached in radiotherapy py. for prostate cancer (see Tables 14 and 15). Does Radical Prostatectomy Improve the Results in Lymph Node-Positive Cases (D1)? 469

Table 15. Outcome of radiation therapy . Groups of patients with node positive and negative disease treated with radiation

No. Survival Free of any failure

5 years 10 years 5 years 10 years

T1B, T2 Node negative 104 87% 63% 85% 67% Node positive 43 60% 24% 38% 20% T3, T4 Node negative 47 82% 58% 69% 49% Node positive 59 65% 26% 32% 10%

Table 16. Comparison of survival rates in cases with and without radical prostatectomy

Stages A and B with pelvic lymph Patients with radical Patients without radical P-value node metastasis D1 prostatectomy prostatectomy Group 1 Group 2 Control without positive nodes 11.2 years 5.8 years 0.005 1±2 positive nodes 20.2 years 5.9 years 0.015 Complication rates 9.5% 24.6%

Because radiotherapy has limited success even The survival advantage of patients who have un- in low-grade cases and in early lymphogenous me- dergone a radical prostatectomy was independent tastatic spread, it seems that even in many patients of any adjuvant therapeutic support. Radical retro- in the early N1 (D1) stages, systemic hematoge- pubic prostatectomy had no perioperative compli- nous spread has already occurred. This has already cations in 72% of the cases. been underlined by the results published by Ger- Complication rates were significantly associated vasi et al. (1989). with anesthesiological problems and comorbidity The risk of distant metastases and of death from rather than age (ASA class P=0.006; operative prostate cancer has been found to be much higher blood loss P = 0.015) (Dillioglugil et al. 1997). in cases with positive nodes than in node-negative Parra et al. (1996) recommend perineal prosta- cases (P<0.00005). tectomy in low-risk cases (PSA <10 ng/ml, Glea- son score <7). In his 75 low-risk cases he found no nodal metastases. In contrast, among 81 cases Does Radical Prostatectomy Improve the Results with worse parameters metastases were detected in in Lymph Node-Positive Cases (D1)? 5 (=6.1%).

The discussion of this question has given rise to some controversy. Opinions on this point vary quite widely and are frequently related to individu- In Conclusion al patients' preconceptions. However, more recent investigations by Frazier · In prostate cancer regional lymph node involve- et al. (1994) give a significant answer to this prob- ment depends heavily on: lem ± Degree of malignancy · With reference to D1 cases and survival ± Stage: whereas in grade I pT1N involvement · With reference to complication rates in cases is nearly zero; in pT2±3 (B2, C) lymph node with and without prostatectomy (see Table 16). involvement is very much more frequent. 470 Chapter 30 Prostate Cancer: an Overview

· When lymph nodes are involved the healing not normally excised during routine lymphade- rates are low, in some statistics near zero. nectomy. · When 99mTc is used for sentinel node detection Even though this approach has made it possible (Wawroschek et al. 1999, 2000), it is possible for to detect atypically localized SLNs, these investiga- atypically localized sentinel nodes to be de- tions must be seen in a critical light, especially tected which would not be detected on open or with reference to the points discussed above. Such laparoscopic lymphadenectomy. atypically localized SLNs can be localized in the · Laparoscopic lymphadenectomy is roughly com- promontory or presacral region (see Fig. 16). parable in value to an open operative procedure, This has some similarities to the situation in but does not allow removal of atypically local- the treatment of rectal cancer. As is well known, in ized SLNs. this cancer the healing rates could be significantly · The chances of achieving more efficient locore- increased by operative excision of the presacral re- gional tumor clearance by an operative proce- gion, thus excluding local recurrence. dure are low, because of the dense networks of New approaches are necessary to clarify lymph and blood vessels (plexus prostaticus, whether it would be possible to inject minimal presacral plexus) in the pelvis, which are limit- amounts of contrast solutions into the capsule re- ing factors. gion, that is to say into regions where the invasive · Immunohistochemical methods (use of cytoke- prostatic cancer is nearest to the capsule or infil- ratin, PSA antibodies) are helpful in the detec- trating the capsule, and to look for the regional tion of micrometastases. lymph nodes that might be involved in a metas- · The following improvements could be helpful to tatic process. It seems to be essential for the vol- increase tumor-free survival: ume of contrast solutions injected to be absolutely ± Development of strongly improved sentinel necessary so as to avoid vascular defects in the in- node detection systems jection region and with these also hematogenous ± Intraoperative fast immunohistochemical metastatic processes in the sense of so-called retro- stainings (Nåhrig, Hæfler) for detection of grade transport via plexus prostaticus prevertebral micrometastases (in sentinel nodes) and ex- plexus into bone marrow of the lumbar column or tended lymphadenectomy depending on the into the ossa ilei (see Fig. 13). result. · Meticulous care to avoid opening venous blood In conclusion, vessels in tumor-infiltrated areas to prevent ret- · Sentinel node labeling in prostatic cancer cases rograde transport via the venous plexus sys- is still in a developmental stage. tems. · The absolutely preliminary results presented by · Improvements of locoregional radiation therapy, the Harzmann group to make it better adapted to the cell cycle of the ± Indicate new ways for operative principles cancer. ± Give an answer on possible atypical met- astatic processes in presacral or other node groups. · According to these reflections and results new Wawroschek's, Vogt's and Harzmann's approaches seem to be possible. First Approach to Detection of Sentinel Nodes · But, the knowledge currently available to us is in Prostate Cancer not adequate to allow recommendations on how to proceed rotuinously in the future. Following on from these open questions, Harz- mann and his group performed a pilot project New approaches are necessary for more differen- with injection of 2 ml of 99mTc containing contrast tiated schemes concerning solution transrectally into the body of the prostate · Site of injection for contrast media. gland. They then faded out the central prostatic re- · Reduction of the amount of fluid injected. gion and looked for the localization of the SLNs We understand from Wawroschek and his collea- by scintigraphic investigations. With this method gues (personal communication) that animal ex- they were able to detect the positions of SLNs out- periments in dogs are in progress to elucidate the side of the iliac and obturator regions, which are facts on the most efficient injection sites, determi- Wawroschek's, Vogt's and Harzmann's First Approach to Detection of Sentinel Nodes in Prostate Cancer 471 nation of the most appropriate volume of solution only the area of node localization can be deter- to be injected and how the total volume injected mined, because there exists no means of imaging might be reduced. the course of the blood vessels at the same time. Nonetheless, in accordance with the sentinel node As things are, it is not possible to be absolutely sure concept only those lymph nodes with measurable that propagation of lymphatic and hematogenous radiation were recognized as sentinels and re- spread is avoided. It is astonishing that the points moved intraoperatively. above are not seriously discussed in all new efforts Subordinate lymph nodes of the particular lym- to developnew techniques. We think a new discus- phatics were excluded from the operative proce- sion should not be started on iatrogenic induction dure. In normal conditions, precise allocation of of hematogenous spread, as this was under discus- subordinate lymph nodes is only successful in the sion as much as 25 years ago, but reflection on the case of nodes along the external and common ilia- problems touched upon does seem to be relevant. cal arteries. Dynamic lymphoscintigraphic investigations As in many histopathology laboratories the were carried out by Harzmann et al. (2000) preop- nodes were cut in 2-mm-thick slices before paraf- eratively on the basis of an informed consent. fin embedding. The gamma probe examinations were per- Besides H and E staining, immunohistochemical formed intraoperatively before pelvic staging lym- studies using antibodies directed to cytokeratin phadenectomy. were also used for detection, especially of single Harzmann's group followed the principle of per- cancer cells or very small cancer cell clusters. forming total-body bone and CT scans, investigat- Preliminary results obtained by Harzmann's ing the acid phosphatase and prostate-specific groupshowed that in a series of patients with antigen levels, and also performing transrectal stage 2A disease (26 cases with grades II and III), ultrasonography, all preoperatively. After these, 4 of 11 cases had 1±2 micrometastases and in 3 of 99mTc nanocolloid (Nanocoll, Sorin, Italy) was 4 cases metastases were found exclusively in the injected into the prostate, one or two injections SLNs. It is remarkable that the largest positive being given into each prostate lobe. When this node measured only 6 mm. The mean number of procedure was followed the total activity reached lymph nodes in pelvic lymphadenectomy was 17.4 100 MBq with a volume of 2 ml. (range 12±20). Harzmann's grouporiented their applications to Vogt et al. (2002) have recently summarized the experiences obtained in breast cancer cases. Augsburg results again. They are collected in Ta- Scintigraphic investigations in the anteroposte- ble 17, which gives an overview of labeled versus rior and dorsal projections were carried out positive SLN, total positive nodes versus positive 12 min and 2±4 h after injection (Sopha camera, SLN, and atypical localizations of positive SLN. DSX, LEAP collimator, 100000 counts/picture). The authors emphasize clearly: Radioactivity of the SLNs was measured using a · Since conspicuous nodes in the presacral and in gamma probe optimized for the measurement of the anteromedial area of the internal iliac artery 99mtechnetium (C-trak: Car-wise Medical Products are not included in the conventional pelvic lym- Co. Morgan Hill, Calif., USA) (see also chap- phadenectomy program, in two of the four pa- ter 12). tients seen by the Harzmann team the microme- As usual in other regimens of sentinel node de- tastases would not have been discovered or ex- tection the prostate gland was shielded by a tung- cised. Had tumor progression developed from sten plate, placed between the prostate and the these nodes an unfavorable outcome would have lymph nodes intraoperatively to inhibit radiation been inevitable, but, the authors clearly point coming from the prostate gland, which could en- out that the more precise surgical procedure de- hance inadequate measurements of activity over mands an operating time of nearly 5 h. the lymph node(s). · Laparoscopic techniques can hardly be used in In their description, the Harzmann group view of this increased duration and of economic explains that a first operative step is the removal calculations. of the nodes detected by preoperative dynamic All in all, the Harzmann concept must be scintigraphy and intraoperative application of the viewed critically because of the large volume of gamma probe. However, the group emphasizes that the contrast solution injected, but on the other 472 Chapter 30 Prostate Cancer: an Overview

Table 17. SLN-positive cases of prostate cancer evaluated by the Augsburg group (study of 25 cases; Vogt et al. 2002)

pT Grade Gleason grade No. of labeled SLN Positive nodes vs Localization of positive LN vs total LN positive SLN pT3b G2b 7 2/22 1/1 F. obt. pT3a G2b 7 3/16 3/3 Ext. iliac., F. obt., presacral pT2v G2b 6 4/18 1/1 Ext. iliac. pT2b G2b 5 5/17 1/1 Int. iliac. pT4 G2b 7 10/36 4/3 F. obt. ext. iliac., int. iliac. pT2b G2b 7 6/18 2/1 F. obt. pT2a G2b 5 3/22 1/1 F. obt. pT2b G2b 7 6/15 4/4 Ext. iliac, int. iliac., presacral pT3a G3a 7 2/13 2/2 Ext. iliac. pT4 G3a 7 4/27 7/4 F. obt. ext. iliac., int. iliac. pT3b G3a 7 2/22 2/2 Ext. iliac., F. obt. pT3a G3a 8 6/24 4/4 F. obt. ext. iliac., int. iliac. pT4 G3a 7 4/26 3/3 Ext. iliac., int. iliac. pT4 G3a 6 2/23 3/2 Int. iliac., ext. iliac. pT3b G3a 8 4/24 3/3 Ext. iliac., F. obt., presacral pT2b G3a 7 5/24 4/3: F. obt. ext. iliac. pT2b G3a 9 1/20 1/1 Int. iliac. pT3b G3a 9 3/11 2/1 F. obt. pT2b G3a 6 6/36 1/1 Int. iliac. pT3b G3a 8 4/12 1/1 Ext. iliac. pT4 G3a 8 6/16 8/6 F. obt. ext. iliac., int. iliac. pT2b G3a 7 8/29 3/3 Ext. iliac., int. iliac. pT3b G3a 9 7/27 6/5 F. obt. int. iliac., ext. iliac. pT3b G3b 9 4/21 4/3 Presacral, int. iliac.

hand the detection of micrometastases is impres- Harzmann's group points out: sively improved and can be of value for the pa- · Preoperative lymphoscintigraphy yields supple- tient. mentary information that is helpful in intra- However, these improvements contrast with the operative gamma probe-guided identification of current situation, according to which laparoscopic SLNs. pelvic lymphadenectomy is performed, allowing · All lymph nodes identified by scintigraphy pre- the detection of micro- and macrometastases in a operatively must be detected by means of the limited range, but a cancer-infiltrated prostate is gamma probe and removed as SLNs. left untreated because curative surgery is not pos- The authors also emphasize that presacral and hy- sible. On the other hand, it must be emphasized pogastric lymph nodes cannot be detected by pre- that radical prostatectomy can minimize local operative scintigraphic investigations because of complications, which develop in more than 20% of the prostate shielding. On the basis of this, it patients treated without curative intent. would be logical to check these nodes as precisely as possible by histological examination. Experiences of the Augsburg Research Group 473

Anatomically it seems to be clear (Fig. 15) that Fig. 15, originally published by Fældi and Kubik following attempts to dissect sentinel lymph nodes 2002). only locally, The schematic illustrations also make it possible · Tumor cells within the lymphatic network in to estimate how the obturator lymph nodes can be the pelvis have a high chance of growing to reached through the so-called obturator window maturity. (see also Fig. 12). · Cancer cells still present locally immediately after breaking through the prostate capsule have a high chance of invading the venous blood vessels, from where they can be retrogradely Experiences of the Augsburg Research Group transported via the prevertebral plexus into the bone marrow of the column (Fig. 13) (Batson In the past few years the members of the Augsburg 1940). Research Groupin Germany have made valuable progress. This section gives an overview of impor- Figure 14 demonstrates the iliac and obturator tant learning effects (see Figs. 15, 16) with refer- lymph node groups, which are the main target ence to the questionable, possibly incomplete, doc- nodes (basins) for metastasis from prostate cancer. umentation of the pelvic lymph node basins. The Readers should be aware that these schematics, grouplooked for more information about drainage mostly depicted in anatomical topographic illustra- from the prostate gland and found the original tions of the pelvic lymphatic-network, do not in- schematic published by Fældi, which clearly docu- clude intercalar (prerectal) lymph nodes, at least ments the direct flow from parts of the prostate some of which are connected with higher located gland to the presacral and deep pelvic nodes presacral and promontorial lymph nodes (see also (Fig. 15).

Fig. 14. a View from left side into pelvis: the iliacal and b View from right side into the pelvis: again the iliacal and obturatoric lymph nodes can be seen in groups (black). obturatoric lymph nodes can be seen in groups (black) 474 Chapter 30 Prostate Cancer: an Overview

Fig. 15. Early documentation of lympha- tics and regional lymph nodes of the ductus deferens, glandulae seminales, and prostate gland (according to Fældi, Germany). The original topographical view of the lymphatic drainage of the prostate gland given by Fældi already notes the possibility of drainage to the presacral lymph nodes (4) and also to the prerectal small lymph nodes (8). The different node groups are indicated (1 lateral lacunar lymph node, 2 inter- mediate external iliac lymph nodes, 3 medial external iliac lymph nodes, 4 promontory common iliac lymph nodes, 5 internal iliac lymph nodes, 6 lateral sacral lymph node, 7 prevesicu- lar lymph node, 8 intercalated nodes)

Experience of pelvic lymph node labeling was 99m growing as a few series of prostate cancer cases Detection of the SLN(s) after Tc-nanocolloid with more than 100 patients in each collective Labeling were studied. Even after finishing their studies on a first series of 117 cases, the groupwas convinced After injection of the labeled nanocolloid solution, that without node labeling the original SLNs are the prostate region is faded out. The regionally lo- left in place in 8±10% of cases (see Fig. 16). In the cated sentinel nodes can be detected by scintigra- top-view photograph shown in Fig. 16 lymph phy as little as 15 min after the injection proce- nodes are marked (black circles) that would have dure and can be clearly seen for 1.5±2.5 h after in- been not excised had the lymph node labeling pro- jection of the labeling solution. cedures not been performed: the rate of missed de- The labeling of pelvic SLNs is demonstrated in tection, meaning non-excision, of lymph nodes is two cases in Figs. 18 and 19. 5% for presacral lymph nodes and 3.4% for the prerectal lymph nodes (see Fig. 16).

Removal of Labeled Lymph Nodes from the Paraprostatic, Iliac, Obturator, Performance of the Labeling Procedure and Retrocolic Basins in Prostate Cancer Cases The main advantages of the new surgical strategies The first stepin the labeling procedure is the ul- with preparation according to the labeling pro- trasound-guided preoperative injection of tracer grams that have been developed are: into the prostate; the status after injection is dem- · The possibility of selecting the best access to onstrated in Fig. 17. reach lymph nodes with suspected involvement (e.g. accessible from the obturator window). Removal of Labeled Lymph Nodes from the Paraprostatic, Iliac, Obturator, and Retrocolic Basins 475

Fig. 16. Pelvic target (sentinel) lymph node groups involved in prostate cancer metastasis. The presacral nodes and dis- tal nodes along the internal iliac artery that are less frequently removed routi- nely though they can be primarily in- volved are indicated by black circles (A external iliac artery, B internal iliac artery, C obturator artery)

· The possibility of removing atypically located lymph nodes, e.g., those in the presacral and deeppelvic region, which otherwise would not be removed. · Optimal monitoring of all putatively involved nodes.

A top-view of an operation site demonstrates a chain of lymph nodes labeled by the SLN-detecting techniques (arrows) though far distant from the iliacal blood vessels; these nodes would not have been included in the ªen bloc-resectionº procedure used for iliacal lymph node excision (Fig. 20). This experience supports the use of the SLN concept for nodal cancer clearance in prostate cancer cases also.

Fig. 17. Prostate after injection of 99mTc-nanocolloid into the cancerous region 476 Chapter 30 Prostate Cancer: an Overview

Fig. 18a,b. Scintigraphic image a 45 min after and b 100 min after injection of 99mTc-nanocolloid solution to a patient with prostate cancer. By 100 min after the injection two supraprostatic lymph node groups are labeled Fig. 19a,b. Labeling of the right-sided iliacal nodes in a case with prostate cancer. One node is sited in a directly para- prostatic position (arrow)(a,b); after 2.5 h two nodes are strongly labeled and can be detected using the gamma probe (b)

Fig. 20. Demonstration of a surgically prepared lymph node chain (white ar- rows) detected by the SLN labeling pro- cedure, at a site far distant from the main-stem of the prepared iliac vessels References 477

localization of the primary. FDG-PET also cannot Is FDG-PET Helpful in Detection and N-staging be used to find cancer-infiltrated regional lymph of Prostate Cancer? nodes in early stages of metastasis. N. Avril, W. Weber, M. Schwaiger

Various studies have shown that FDG is not suitable Can Sentinel Node Labeling be Improved for diagnosing changes in the prostate gland, as According to Animal Experimental Studies? prostate cancer is often not accompanied by an in- crease in glucose metabolism. In a study of primary The current international literature and, especially, prostate cancer, Effert et al. (1996) noted only low the experience and knowledge available within the metabolic activity in most tumors, which was not Augsburg Group, current practice needs further related to tumor grade or stage. Data from 11 pa- improvement. tients with localized prostate cancer and 2 with be- The intention is now to use the lowest possible nign prostate hyperplasia confirmed these results dosages of radioactivity and the smallest possible (Hofer et al. 1999). Following radical prostatectomy injection volumes, with the total volume to be in- and an increase in prostate-specific antigen (PSA), it jected split into small portions for multifocal injec- is not possible to differentiate between scar tissue tion into the prostate parenchyma. Injection of the and local recurrence by means of FDG-PET. Of 6 pa- labeling solution should be ultrasound guided. It tients with a local recurrence diagnosed by biopsy, 5 is important to avoid injection into the venous ves- had false-negative results in the PET scan (Hase- sels of the prostate plexus. mann et al. 1996). At the same time, 2 out of 4 pa- These points are of basic importance for the de- tients with negative biopsies showed enhanced velopments of experimental research plans. New FDG-uptake. Therefore, metabolic activity of local approaches should be developed further and tested recurrences following radical prostatectomy cannot in dogs. be distinguished from vascularized metabolic active scar tissue by FDG-PET. In a study of 34 patients with metastasized prostate cancer, Shreve et al. (1996) examined the value of FDG-PET for the diag- References nosis of bone metastases, using CT, bone scintigra- phy, and clinical follow-up as a reference. Among Abbas F, Scardino T (1993) Why neoadjuvant androgen de- the 22 untreated patients there were 202 bone metas- privation prior to radical prostatectomy is unnecessary. Urol Clin North Am 23:587±604 tases in all, only 131 of which were detected using Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane FDG-PET. In the case of small metastases, the low CG, Epstein JI (2001) Interobserver reproducibility of sensitivity of 65% is due at least in part to partial Gleason grading of prostatic : general patholo- volume effects. The most important cause, however, gists. Hum Pathol 32(1):81±88 is the low metabolic activity and the resultant low Altay B, Kefi A, Nazli O, Killi R, Semerci B, Akar I (2001) Comparison of Gleason scores from sextant prostate FDG accumulation of prostate tumors and their me- biopsies and radical prostatectomy specimens. Urol Int tastases. Despite this, a high positive-predictive re- 67(1):14±18 sult of 98% was obtained for metabolically active le- Arduino LJ, Glucksman MA (1962) Lymph node metastases sions. The results obtained by Shreve et al. have in early carcinoma of the prostate. J Urol 88:91 Barzell W, Bean MA, Hilaris BS, Whitmore WF Jr (1977) been confirmed by other studies conducted by smal- Prostatic adenocarcinoma: relationshipof grade and ler groups. In a comparison with positive foci iden- local extent to the pattern of metastases. J Urol Vol tified using skeletal scintigraphy, Yeh et al. (1996) 118:278±282 found only some 18% of the lesions in 13 patients Batson OV (1940) Function of vertebral veins and their role with pretreated bony prostate metastases, which in spread of metastases. Ann Surg 112:138 Berner A, Waere H, Nesland JM, Paus E, Danielsen HE, Fos- had previously been classified as refractory to hor- sa SD (1995) DNA ªploidyº serum prostate specific anti- mone treatment. gen, histological grade and as

'In summary, in most cases of prostate cancer predictive parameters of lymph node metastases in T1± FDG-PET is not helpful in the detection and delin- T3 M0 prostatic adenocarcinoma. Br J Urol 75:26±32 eation of prostatic primaries, which means it also Beurton D, Barthlmy Y, Fontaine E (1997) Twelve sys- tematic prostate biopsies are superior to sectant biopsies has no value for the SLN concept, because sentinel for diagnosing carcinoma: a prospective randomized node localization depends at least in part on the study. Br J Urol 80:239 478 Chapter 30 Prostate Cancer: an Overview

Bluestein DL, Bostwick DG, Bergstralh EJ, Oesterling JE Frazier HA II, Robertson JE, Paulson DF (1994) Does radi- (1994) Eliminating the need for bilateral pelvic lympha- cal prostatectomy in the presence of positive pelvic denectomy in select patients with prostate cancer. J Urol lymph nodes enhance survival. World J Urol 12:308±312 151:1315±1320 Freiha FS, Salzman J (1977) Surgical staging of prostatic Bostwick DG, Eble JN (1993) Prostatic adenocarcinoma me- cancer: transperitoneal versus extraperitoneal lymphade- tastatic to inguinal hernia sac. J Urol Pathol 1:193 nectomy. J Urol 118:616 Brendler CB, Cleeve LK, Anderson EE , Paulson DF (1980) Freiha FS, Pistenma DA, Bagshaw MA (1979) Pelvic lympha- Staging pelvic lymphadenectomy for carcinoma of the denectomy for staging prostatic carcinoma: is it always prostate: risk versus benefit. J Urol 124:849 necessary? J Urol 122:176 Bruce AW, O'Cleireachain F, Morales A, Awas SA (1977) Car- Gervasi LA, Mata J, Easley JD, Wilbanks JH, Seale-Hawkins cinoma of the prostate: a critical look at staging. J Urol C, Carlton CE Jr, Scardino PT (1989) Prognostic signifi- 117:319 cance of lymph nodal metastases in prostate cancer. J Carlson GD, Calvanese CB, Kahane H, Epstein JI (1998) Ac- Urol 142:332 curacy of biopsy Gleason score from a large uropathol- Gomella LG, White JL, McCue PA, Byrne DS, Mulholland ogy laboratory: use of a diagnostic protocol to minimize SG (1993) Screening for occult nodal metastasis in local- observer variability. 51(4):525±529 ized carcinoma of the prostate. J Urol 149:776±778 Catalona WJ, Stein AJ (1982) Staging errors in clinically lo- Gregori A, Vieweg J, Dahm P, Paulson DF (2001) Compari- calized prostatic cancer. J Urol 127:452 son of ultrasound-guided biopsies and prostatectomy Catalona WJ, Smith DS (1994) 5-year tumor recurrence specimens: predictive accuracy of Gleason score and tu- rates after anatomical radical retropubic prostatectomy mor site. Urol Int 66(2):66±71 for prostate cancer. J Urol 152:1837±1842 Grossman IC, Carpinielle V, Greenberg SH, Malloy TR, Wein Cheng CWS, Bergstralh EJ, Zincke H (1993) Stage D1 pros- AJ (1980) Staging pelvic lymphadenectomy for carcino- tate cancer: a non randomized comparison of conserva- ma of the prostate: review of 91 cases. J Urol 124:632 tive treatment options versus radical prostatectomy. Can- Hanks GE (1993) The challenge of treating node positive cer Suppl 71(3):996±1004 prostate cancer. An approach to resolving the questions. Conrad St, Graefen M, Pichlmeier U, Henke RP, Hammerer Cancer 71 [3 Suppl]:1014±1018 PG, Huland H (1998) Systematic sextant biopsies im- Haseman MK, Reed NL, Rosenthal SA (1996) Monoclonal prove preoperative prediction of pelvic lymph node me- antibody imaging of occult prostate cancer in patients tastases in patients with clinically localized prostatic car- with elevated prostate-specific antigen. Positron emission cinoma. J Urol 159:2023±2029 tomography and biopsy correlation. Clin Nucl Med Danella JF, Kernion JB de, Smith RB, Steckel J (1993) The 21:704±713 contemporary incidence of lymph node metastases in Hofer C, Laubenbacher C, Block T et al (1999) Fluorine-18- prostate cancer: implications for laparoscopic lymph fluorodeoxyglucose positron emission tomography is node dissection. J Urol 149:1488±1491 useless for the detection of local recurrence after radical Davies GL (1995) Sensitivity of frozen section examination prostatectomy. Eur Urol 36:31±35 of pelvic lymph nodes for metastatic prostate cancer. Huland H (1998) Welchen Stellenwert hat die radikale Pros- Cancer 75:662±668 tatektomie beim Lymphknoten positiven Prostatacarci- Dillioglugil O, Leibman BD, Leibman NS, Kattan MW, Rosas nom? Urologe 37:138±140 AL, Scardino PT (1997) Risk factors for complications King CR, Long JP (2000) grading errors: a and morbidity after radical retropubic prostatectomy. J sampling problem? Int J Cancer 90(6):326±330 Urol 157:1760±1767 Koksal IT, Ozcan F, Kadioglu TC, Esen T, Kilicaslan I, Tunc Donohue RE, Mani JH, Whitesel JA, Mohr S, Scanavino D, M (2000) Discrepancy between Gleason score of biopsy Augspurger RR, Biber RJ, Fauver HE, Wettlaufer JN, Pfis- and radical prostatectomy specimens. Eur Urol 37(6): ter RR (1982) Pelvic lymph node dissection. Guide to 670±674 patient management in clinically locally confined adeno- Kozlowski JM, Grayhack JT (1987) Carcinoma of the pros- carcinoma of prostate. Urology 20:559 tate. In: Gillenwater JY, Grayhack JT, Howard SS, Duckett Effert PJ, Bares R, Handt S, Wolf JM, Bull U, Jakse G (1996) JW (eds) Adult and pediatric urology, vol 2, chap 34. Metabolic imaging of untreated prostate cancer by posi- Year Book Medical Publishers, Chicago, pp 1126±1130 tron emission tomography with 18-fluorine-labeled deox- Kramer STA, Cline WA Jr, Farnham R, Carson CC, Cox EB, yglucose. J Urol 155:994±998 Hinshaw W, Paulson DF (1981) Prognosis of patients Epstein JI, Oesterling JE, Eggleston JC, Walsh PC (1986) with stage D1 prostatic adenocarcinoma. J Urol 125:817± Frozen section detection of lymph node metastases in 819 prostatic carcinoma: accuracy in grossly uninvolved pel- Leibovitch I, Westenfelder K, Vaught J, Rowland RG (1995) vic lymphadenectomy specimens. J Urol 136:1234±1237 Orthostatic abnormal penile erections: a consequence of Eskew LA, Bare RL, McCullough DL (1996) Systematic 5 re- retroperitoneal lymphadenectomy with vena caval resec- gion prostate biopsy is superior to sextant method for tion. J Urol 154:533±534 diagnosing carcinoma of the prostate. J Urol 157:199 Lieskovsky G, Skinner DG (1983) Technique of radical ret- Flocks RH, Culp D, Porto R (1959) Lymphatic spread from ropubic prostatectomy with limited pelvic node dissec- prostatic cancer. J Urol 81:194 tion. Urol Clin N Am 10:187 Fældi M, Kubik S (2002) Lehrbuch der Lymphologie, 5th Lilleby W, Torlakovic G, Torlakovic E, Skovlund E, Fossa SD edn, p150. Urban and Fischer, Munich (2001) Prognostic significance of histologic grading in Fowler JE Jr, Whitmore WF Jr (1981) The incidence and ex- patients with prostate carcinoma who are assessed by tent of pelvic lymph node metastases in apparently local- the Gleason and World Health Organization grading sys- ized prostatic cancer. Cancer 57:2941 tems in needle biopsies obtained prior to radiotherapy. Cancer 92:311±319 References 479

Loening SA, Schmidt JD, Brown RC, Hawtrey CE, Fallon B, Paulson DF (1989) Remark to the publication of Gervasi et CulpDA (1977) A comparison between lymphangiogra- al. J Urol 142:336 phy and pelvic node dissection in the staging of pros- Petros JA, Catalona WJ (1992) Lower incidence of unsus- tatic cancer. J Urol 117:752 pected lymph node metastases in 521 consecutive pa- McCarthy P, Pollak HM (1991) Imaging of patients with tients with clinically localized prostate cancer. J Urol stage D prostatic carcinoma. Urol Clin N Am 18:35 147:1574±1575 McCullough DL, McLaughlin AP, Gittes RF (1977) Morbidity Ravery V, Billeboud T, Toublanc M, Boccon-Gibod L, Her- of pelvic lymphadenectomy and radical prostatectomy mieu JF, Moulinier F, Blank E, Delmas V, Boccon-Gibod for prostatic cancer. J Urol 117:206 L (1999) Diagnostic value of ten systematic TRUS-guided McCullough DL, Prout GR Jr, Daly JJ (1974) Carcinoma of prostate biopsies. Eur Urol 35:298 the prostate and lymphatic metastases. J Urol 111:65 Schçssler WW, Pharaud D, Caillie TG van (1993) Laparo- McDowell GC II, Johnson JW, Tenney DM, Johnson DE scopic standard pelvic node dissection for carcinoma of (1990) Pelvic lymphadenectomy for staging clinically lo- the prostate: is it accurate. J Urol 150:898±901 calized prostatic cancer. Indications, complications, and See WA, Cohen MB, Winfield HN (1993) Inverted V-perito- results in 217 cases. Urology 35:476 neotomy significantly improves nodal yield in laparo- McLaughlin AP, Saltzstein SL, McCullough DL, Gittes RF scopic pelvic lymphadenectomy. J Urol 149:772±775 (1976) Prostatic carcinoma: incidence and location of Shreve PD, Grossman HB, Gross MD, Wahl RL (1996) Meta- unsuspected lymphatic metastases. J Urol 115:89 static prostate cancer: initial findings of PET with 2- McNeal JE, Villers A, Redwine EA, Freiha FS, Stamey TA deoxy-2-[F-18]fluoro-D-glucose. 199:751±756 (1990) Histologic differentiation, cancer volume and pel- Smith JA Jr, Seaman JP, Gleidman JB, Middleton RG (1983) vic lymph node metastases in adenocarcinoma of the Pelvic lymph node metastasis from prostatic cancer: in- prostate. Cancer 66:1225±1233 fluence of tumor grade and stage in 452 consecutive pa- McNeal JE (1993) Prostatic microcarcinomas in relation to tients. J Urol 130:290±292 cancer origin and the evolution to clinical cancer. Cancer Stamey TA, Freiha FS, McNeal JE, Redwine EA, Whitmore (Suppl 3) 71:984±991 AS, Schmid HP (1993) Localized prostate cancer: rela- Moul JW, Lewis DJ, Ross AA, Kahn DG, Ho CH, McLeod DG tionshipof tumor volume to clinical significance for (1994) Immunohistologic detection of prostate cancer pel- treatment of prostate cancer. Cancer 71(Suppl 3):933±938 vic lymph node micrometastases correlation to preopera- Vogt H, Wawroschek F, Wengenmair H, Wagner T, Kopp J, tive serum prostate specific antigen. Urology 43:68 Dorn R, Græber S, Heidenreich P (2002) Sentinel lymph Naughton CK, Miller DC, Mager DE, Ornstein DK, Catalona node diagnostic in prostate carcinoma. Technique and WJ (2000) A prospective randomized trial comparing 6 clinical evaluation. Nuklearmedizin 41:95±101 versus 12 prostate biopsy cores: Impact on cancer detec- Walsh PC, Partin AW, Epstein JI (1994) Cancer control and tion. J Urol 164:388±392 quality of life following anatomical radical retropubic Nicholson TC, Richie JP (1977) Pelvic lymphadenectomy for prostatectomy: results at 10 years. J Urol 152:1831±1836 stage B1 adenocarcinoma of the prostate: justified or Wawroschek F, Vogt H, Weckermann D, Wagner T, Harz- not? J Urol 117(2):199±201 mann R (1999) The sentinel lymph node concept in Oesterling JE, Chan DW, Epstein JI, Kimball AW Jr, Brucek prostate cancer ± first results of gamma probe-guided DJ, Rock RC, Brendler CB, Walsh PC (1988) Prostate-spe- sentinel lymph node identification. Eur Urol 36:595±600 cific antigen in the preoperative evaluation of localized Wawroschek F, Vogt H, Bachter D, Weckermann D, Hamm prostatic cancer treated with radical prostatectomy. J M, Harzmann R (2000) First experience with gamma Urol 139:766±772 probe guided sentinel lymph node surgery in penile can- Okegawa T, Nutahara K, Higashihara E (2000) Detection of cer. Urol Res 28:246±249 micrometastatic prostate cancer cells in the lymph nodes Weingårtner K, Ramaswamy A, Bittinger A, Gerharz EW, by reverse transcriptase polymerize chain reaction is Dæge D, Riedmiller H (1996) Anatomical basis for pelvic predictive of biochemical recurrence in pathological lymphadenectomy in prostate cancer: results of an stage T2 prostate cancer. J Urol 163:1183±1188 autopsy study and implications for the clinic. J Urol 156: Parra RO, Andrus C, Boullier J (1992) Staging laparoscopic 1969±1971 pelvic lymph node dissection comparison of results with Whitmore WF Jr (1984) Natural history and staging of open pelvic lymphadenectomy. J Urol 157:875±878 prostate cancer. Urol Clin N Am 11:205 Parra RO, Isorna S, Perez MG, Cummings JM, Boullier JA Wilson CS, Dahl DS, Middleton RG (1977) Pelvic lymphade- (1996) Radical perineal prostatectomy without pelvic nectomy for the staging of apparently localized prostatic lymphadenectomy: selection criteria and early results. J cancer. J Urol 117:197 Urol 155:612±615 Winfield HN, Kavoussi LR (1991) Laparoscopic pelvic node Partin AW, Carter HB, Chan DW, Epstein JI, Oesterling JE, dissection. Surg Tech Urol 4:1 Rock RC, Weber JP, Walsh PC (1990) Prostate specific Winfield HN, Donovan JF, See WA, Loening SA, Williams antigen in the staging of localized prostate cancer: influ- RD (1991) Urological laparoscopic surgery. J Urol 146: ence of tumor differentiation, tumor volume and benign 941 hyperplasia. J Urol 143:747±752 Yeh SD, Imbriaco M, Larson SM, Garza D, Zhang JJ, Kalai- Paulson DF (1980) The prognostic role of lymphadenectomy gian H, Finn RD, Reddy D, Horowitz SM, Goldsmith SJ, in adenocarcinoma of the prostate. Urol Clin N Am Scher HI (1996) Detection of bony metastases of andro- 7:615 gen-independent prostate cancer by PET-FDG. Nucl Med Biol 23:693±697