Recent Advances in Gastrointestinal Cancer

1 GALEN L. WAMPLER, MD

Associate Professor of Medicine, Division of Medical Oncology, Department of Medicine, Medical College of Virginia, Health Sciences Division of Virginia Commonwealth University, Richmond, Virginia

INTRODUCTION modest, and for this reason the drugs have Colorectal carcinoma accounts for the been combined into a variety of multiple-drug majority of all gastrointestinal cancers and is the regimens. Most of the possible two and three­ second leading site of cancer, excluding skin drug combinations of the four active drugs have cancers, in overall incidence in the United been tried. States. 1 Cancer of the stomach, although de­ Table 1 shows response rates and sur­ creasing in frequency, is still an important cause vival figures for some of the most extensively of morbidity and mortality. Unfortunately, data tested combinations. The combination of 5-FU from large numbers of patients such as can be and semustine was one of the first advocated as found in Cancer Patient Survival Report No. 5 being superior to 5-FU alone in the treatment of show only very modest increases in survival for gastric carcinoma. More recently 5-FU plus patients with th\;!Se diseases in recent years. 2 Adriamycin plus mitomycin-C (FAM) combina­ tions have become more popular. The FAM Gastric Carcinoma regimen was initially reported to have a 50% re­ Most patients who develop gastric cancer sponse rate in gastric carcinoma. The most re­ have regional or distant disease at diagnosis. cent update of over 60 patients indicates that The presence of regional nodal involvement is the response rate is holding at approximately almost synonymous with incurability since virtu­ 43%. 3 ally all these patients are either non-resectable The FAM regimen (Table 2) is a well-toler­ at the tim(;l of operation or rapidly develop sys­ ated treatment which gives partial or complete temic recurrences. Any improvement in treat­ responses in about one half of the patients and ment results would be expected to be produced benefits other patients by stabilizing the dis­ by chemotherapy or immunotherapy rather than ease, resulting in prolonged survival for the pop­ from localized forms 9f treatment. ulation of treated patients. Quality of survival for Fortunately, gastric carcinoma is relatively many is good, and it is not uncommon to see responsive to chemotherapeutic treatment, and responses lasting for over one year. at least four drugs have now been identified as One problem with the treatment is cu­ active in treatment of this condition, namely 5- mulative marrow toxicity which is attributed to fluorouracil (5-FU), Adriamycin, mitomycin-C, the mitomycin-C in the regimen . This tends to and semustine (methyl-CCNU). Although se­ limit the treatment that can be given after the mustine is the nitrosourea that has been most first few cycles. The cycles of treatment are sim­ extensively used in the chemotherapy of gastro­ ilar to other day-1 , day-8 treatments given every intestinal neoplasms, it is still an investigational eight weeks. drug and therefore is not always conveniently Other combinations of these drugs which available. Other nitrosoureas that are on the utilize different doses and regimens have also market are probably similar in activity. been tried. Two of these used at the Sloan Ket­ Percentage of response and increase in tering Institute known as MIFA I and MIFA II con~ survival with single-drug therapy have been firm that these drugs in combination are effec-

90 / WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER TABLE 1 Combinations Used in Treatment of Gastric Carcinoma

Response Rate Med. Survival Weeks

5-FLUOROURACIL + Semustihe 9-45% 17-25 5-FLUOROURACIL + ADRIAMYCIN + MITOMYCIN-C 21-43% 24-34 5-FLUOROURACIL + ADRIAMYCIN + Semustine 36% 22-30 5-FLUOROURACIL + MITOMYCIN-C 17 ADRIAMYCIN + MITOMYCIN-C

tive in the treatment of gastric cancer. 4 Table 1 operative ahd postoperative radiation therapy indicates treatment results of these and other given adjuvantly in high-risk patients would be combination treatments for gastric cancer. The beneficial , particularly in patients with carci­ median survival of patients with untreated ad­ noma of the rectum, these suggestions have not vanced gastric carcinoma is four months or met with widespread acceptance. about 1 7 weeks from diagnosis. Since the fig­ Five-fluorouracil has been a standard ures in Table 1 show the survival in weeks from treatment for colorectal cancer for 20 years , the time of treatment rather than diagnosis for and one might summarize the clinical experi­ the entire population of treated patients, not just ence of a number of investigators' attempts to the responders, one can see that a doubling of improve results by manipulating the 5-FU dose, the survival time for the better combination is schedule, or route of administration by simply achievable. stating that no schedule of treatment has been Dr Charles Moertel from the Mayo Clinic definitely shown to be superior to any other. 6 has recently analyzed data combined from sev­ The most common schedules used have been eral cooperative groups. Using a statistical daily intravenous. treatments for five days re­ model, he concluded that 5-FU and Adriamycin peated at five-week intervals or one intravenous contribute most to the treatment of gastric can­ treatment administered weekly. cer. 5 This is a combination that has not received Administered orally, 5-FU gives response extensive use, and the projected value of the rates similar to the intravenous treatments of the treatment needs confirmation in a large clinical drug . However, Moertel 7 has reported that the trial. duration of response is shorter with the oral form of treatment. Absorption is erratic, averag­ Colorectal Carcinoma ing about 50%. Because of these facts and be­ Although there have been a number of at­ cause no oral form of treatment has been mar­ tempts to improve results in treatment of colo­ keted, use of this drug by this route of rectal cancer, most of these have been futile or administration has not gained wide acceptance. have achieved only very modest success. The Response rates with 5-FU in colorectal surgical treatment of colorectal cancer has carcinoma average approximately 20%. Two been standard for several decades, and al­ forms of therapy which have a response rate though some recent studies indicate that pre- above 1 0% are the nitrosoureas and mitomy­ cin-C . Other drugs have either had limited use in colorectal cancer or have given response TABLE 2 rates of 1 0% or less, leaving only a few drugs FAM Regimen that have a significant response rate in this dis­

DAY 1 DAY 8 DAY 29 DAY 36 ease. F F F F A number of combinations have been de­ A A vised for the treatment of colorectal carcinoma, M the more extensively tested combinations being : 5-FU with semustine, mitomycin-C or hy­ F = 5-Fluorouracil 600 mg/M2 droxyurea; 5-FU plus semustine plus vincristine; A= Adriamycin 30 mg/M2 M = Mitomycin-C 1 O mg/M2 and 5-FU plus semustine plus dacarbazine. Re­ Cycles of treatment are repeated every eight weeks. sponse rates for the combinations were initially

WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER / 91 reported to exceed the results of 5-FU alone. FU in adjuvant treatment. In the non-random­ The combination of 5-FU, semustine and vin­ ized studies using historical controls, 5-FU was cristine has been said to have response rates in reported to produce a beneficial effect 13 · 14 ; in the range of 35 to 40% by at least three differ­ the randomized studies the 5-FU , in all cases, ent groups. 8 •9 · 10 However, as additional studies produced a slight prolongation of disea$e-free and survival data are reported, the superiority of interval and survival in the treated group. 15- 19 this combination over 5-FU alone has not been lnitialiy, the difference was judged not to be sta­ confirmed .6 tistically significant; however, a recent statistical Median survival for patients after proof of analysis using cumulative results involving incurability is approximately 30 to 32 weeks larger patient numbers resulted in the con­ with 5-FU alone, anq for the combinations the clusion that there is a statistically significant im­ survival has be.en in the same range . Con­ provement, at least for some subsets of 5-FU­ sequently, the current consensus is that no treated patients in the adjuvant setting. Results combination of drugs for the treatment of colo­ of these studies are still pending . rectal carcinoma has proved to be superior to 5- In an article in the Annals of Internal Medi­ FU alone. At this time other combinations are cine, 20 Ors Weiss and Devita stated that at the being tried which , it is hoped, will yield results present time whether or not a patient receives surpassing those with only 5--FU. adjuvant chemotherapy for colorectal carci­ There is controversy regarding whether or noma is a decision that needs to be individ­ not 5-FU alone increases survival in patients ualized for each patient. This is primarily be­ with colorectal carcinoma. To my knowledge, cause the results of 5-FU have been marginal at no prospective randomized trials have been best, and the potential benefits of therapy may done comparing 5-FU with no treatment in pa­ be overridden by a variety of other factors: dis­ tients with advanced disease. Used many years ease stage (patients with Duke 's 82 or C stage ago for the treatment of advanced colorectal lesions, eg, extension through the muscular cancer, 5-FU was shown to yield responses and layer and/ or invoived nodes are customarily was felt to be beneficial , not only in achieving treated); age of the patient; histologic grade of these responses but also in extenc;jing the life of the tumor; economic factors; and convenience the patient. No 9ne has since been willing to of travel to the treatment center for the patient. compare it to no treatment. Following a trial of adjuvant treatment, the deci­ One can easily demonstrate that 5-FU re­ sion to continue treatment should be based on sponders live longer than non-responders. Ad­ the patient's tolerance tempered by the knowl­ ditionally, retrospective analysis indicates that edge of limited survival benefit. 5-FU produces a modest increase in median survival in a population of treated patients. Other Gas\rointestinal Tumors Moertel' s own data 1 1 demonstrated that 5-FU There are three uncommon types of gas­ treated patients live longer throughout the entire trointestinal malignancies, all of which show sig­ survival curve than matched historical controls. nificant response rates to treatment with chemo­ This difference was discounted by Moertel who therapy. stated that there is no evidence showing that 5- Leiomyosarcomas are found in the stom­ FU prolongs survival, attributing the difference ach and in the bowel . Recurrent or metastatic to patient selection. This is only an opinion, and tumors respond to treatment with Adriamycin in different interpretations are possible. approximately 30% of cases, and if combined A number of surgical adjuvant trials have with dacarbazine, the response rate may be been conducted in patients with colon and rec­ 1 0% higher. The treatment for metastatic tal carcinomas. One of the earliest studies uti­ leiomyosarcomas of the bowel is the same as lized thiotepa or fluorodeoxyuridine (FUdR) after for other metastatic sarcomas. surgery. This particular study showed no effect. Lymphomas also occur in the gastrointes­ However, it is of interest that these patients, fol­ tinal tract, and although their natural history lowed over a decade, had no increase in carci­ may be somewhat different from those origina­ nogenicity or other lc;3.te toxicity which could be ting els~where, the chemotherapeutic treatment ascribed to those treatments. 12 More recentiy a is basically similar. Where the histology predicts number of studies have been done utilizing 5- a favorable outcome, treatment would probably

92 / WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER consist of cyclophosphamide plus vincristine found to have elevated CEA levels above 5 ng / plus prednisone (COP) therapy and the unfavor­ ml. Workup of these 73 patients resulted in the able ones treated in addition with Adriamycin finding of malignancy in only 11 patients. Nine and possibly with Bleomycin. had a CEA-related malignancy and two had an Carcinoid tumors are also responsive to incidental malignancy. The false-positive rate , chemotherapeutic treatment. About one third of therefore, was calculated at 87%. these patients with advanced disease show ob­ More disturbing than even the high false­ jective responses to treatments with 5-FU and positive rate was the fact that 1 6 patients who streptozotocin in combination. More recently were CEA-negative developed a CEA-associ­ Adriamycin has been used as a single agent. ated cancer during the follow-up period for a Treatment for this condition is still very much in false-negative rate of 64%. Only 3% of the a stage of evolution. 2, 3 7 2 patients had elevated CEA levels which is consistent with the fact that 95% of a normal Other Advances population are known to have CEA below 2. 5 The biologic marker known as carcinoem­ ng/ ml. (In this population, 97% had the CEA bryonic antigen (CEA) which has been devel­ level below 5 ng/ ml.) The low incidence of can­ oped for clinical use in the last decade, ~as cer in this population of 2 , 3 7 2 resulted in the contributed materially to our ability to stage and high false-negative percentage. CEA testing follow patients with colorectal carcinoma. The may play a role in screening certain high-risk initial hope was that the test would be useful as populations, but it is not suitable as a screening a diagnostic and screening tool. It has not mechanism for carcinoma of the colon in unse­ proved to be very useful for this purpose. How­ lected populations. ever, it has been found to be beneficial as a It was originally thought that CEA would prognostic indicator. Patients with high levels of be specific for colon carcinoma since the anti­ CEA prior to surgery will not do as well as pa­ gen was obtained from fetal colonic ti ssue. tients with normal levels. It can also be utilized However, it was soon found that it is non-spe­ to assess adequacy of treatment or to evaluate cific for colon carcinoma, being elevated in a disease recurrence and treatment response. It variety of other malignancies including breast, has been suggested that CEA-producing tu­ lung, pancreas, stomach and bladder carcino­ mors are inherently more likely to metastasize mas, and in other malignancies. CEA is also and are less controllable by the body's immune elevated in patients with liver disease, pancre­ processes. This biological difference, if con­ atic cysts, gastrointestinal polyps and other be­ firmed, will undoubtedly influence future treat­ nign conditions. The CEA is not specific for a ment strategies. particular primary site and is not even as spe­ Table 3 presents data taken from a study cific as one would like for malignancy. where 2, 3 7 2 patients of an unselected popu­ With the understanding of certain charac­ lation were screened for malignancy using CEA teristics of biologic markers shown in Table 4, values. 2 1 Seventy-three of these patients were CEA tests can be used quite advantageously for a number of purposes. In general, biologic markers are non-specific for histologic types of TABLE 3 Use of CEA In Screening an Unselected Populatlon21 neoplasms and often are not even specific for 23 72 people followed 5 years malignancy. The specificity and sensitivity tend 87% False Positive 64% False Negative to be inversely related . By developing a more CEA LEVEL sensitive assay, more positive results will be ob­ < 5 ng/ml ::::5 ng / ml tained in patients with non-malignant conditions Developed a 16 11* 25 or with tumor types other than those antici­ CEA related Cancer pated. In contrast, if normal levels are drawn at Never developed 2000+ 62 a higher concentration, the test is more specific . Cancer For example, most patients with CEA over 1 0 ng/ml will have a malignancy. 22 The percent­ 73 age of patients with positive markers increases (3 % of total population) • two others were found to have incidental cancers not with the stage of the disease. Reports of CEA related to CEA evaluation. elevation in the 90% range are applicable only

WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER / 93 TABLE 4 less prior to recurrence. The percentage of pa­ Some Characteristics of Biologic Markers tients who show elevated CEA prior to recur­ 1 . In general biologic markers are non-specific for a rence of colon cancer is higher than in rectal histologic type of neoplasm and often are not even specific for malignancy. cancer. (Table 5). 2. Sensitivity and specificity tend to be inversely related. In colon cancer only 14% are never ele­ 3. The percentage of patients with positive markers vated prior to documentation of recurrence increases with stage of disease. compared to 32% for rectal cancer. While this 4. Not all patients develop positive markers. looks quite good as a tool for predicting recur­ 5. Marker status is not a dependent variable in relatipn to staging. rences, it has to be tempered by the fact that matched controls also have a high percentage of at least one elevated CEA value. The to patients with advanced colorectal carcinoma. matched controls were patients with the same The figure is much lower for patients with local­ age, the same disease, and treatment, who had ized or regional disease only. not had a recurrence during periods of equal Not all patients develop positive markers. follow-up. Some of these patients will eventually Only those patients with certain phenotypic can­ turn out to have a recurrence because it is cer cell expressions will show marker elevation . known that the CEA can be elevated for as long Other patients having histologically-similar tu­ as several years prior to recurrence. Some of mors will never be marker positive. Therefore, it these patients have random increases in CEA is futile to attempt to manipulate a test to give value; others have benign causes of the eleva­ results 1 00% of the time or to look for new tion. markers that will do this. It is known that patients who receive blood The percentage of CEA positivity in pa­ products at the time of their surgery sometimes tients with colonic cancer was initially reported develop CEA elevations that plateau and later in excess of 90%. Later the percentage fell, the decrease2 5 secondary to hepatitis or undeter­ reason being that initially patients with ad­ mined factors in the absence of acute or vanced disease were tested , and in the later se­ chronic liver disease. The problem in following ries more patients with earlier stages of disease these patients is to separate those who have were tested .2 3 The percentage of CEA elevation random elevations or benign conditions from directly correlates with the stage of disease; those who have a recurrence of malignancy. In however, it is not a dependent variable, mean­ order to distinguish those with random eleva­ ing that the distribution of CEA positives in pa­ tions, a CEA nomogram has been developed to tients of various stages has a tendency toward indicate when values are statistically increased. randomness. In general, CEA and other bio­ Nomograms are published in the literature, 26·2 7 logic markers are not dependent variables in re­ but to be valid , each laboratory should con­ lation to the stage of disease or any other struct its own based on the precision of its test. known prognostic factor. This means that prog­ While the nomogram is useful in differentiating nostication is improved by considering marker patients who have a random increase from values along with stage, grade of tumor and those with a true increase in CEA levels, repeat­ other standard prognostic indicators. ing the value several times helps in making this CEA is also useful in following colorectal differentiation. patients for recurrence. The majority of patients It is important to make a diagnosis of re­ who are found to have recurrences will have currence earlier using CEA elevations. Some of shown at least one CEA elevation greater than these patients may be candidates for second 2.5 ng/ml more than three months before doc­ surgery; others will be candidates for radiation umentation of recurrence. One study2 4 showed therapy or chemotherapy. In considering pos­ 54% of the patients having this marker positive sible patients for second surgery, it is necessary (>2.5 ng/ml) more than three months prior to to separate those patients who are going to the documentation of recurrence. If a higher have an operable malignancy from those who positive value (5 ng/ml) is used , 41 % will have have random fluctuation of the CEA, benign an elevation three months before clinical tumor conditions, and inoperable lesions. Using the recurrence. An additional number of patients nomogram, one can distinguish most random will have markers positive for three months or fluctuations. Benign conditions tend to be

94 / WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER TABLE 5 Percentage of Patients With Established Tumor Recurrence Who Exhibit CEA Elevations 24

Primary >2.5 ng / ml >5 ng/ml >2.5 ng / ml >5 ng / ml Tumor >3 mos before >3 mos before <3 mos before <3 mos before Never Site recurrence recurrence recurrence recurrence elevated

COLON 58 45 28 26 14 RECTUM 42 31 26 32 32 COMBINED 54 41 29 28 17 nomogram positive as are both operable and in­ A factor to be considered in the decision operable malignancies. The operable malig­ for reoperability is the interval between surgery nancies and random fluctuations would not be and observed CEA elevation . The benign causes expected to have elevated liver enzymes. The of the CEA elevations occur earlier after surgery degree of CEA elevation helps to discern the than the operable malignancies.21 Any CEA ele­ operable and inoperable malignancies. High vations that occur early are more likely to be as­ elevations correlate with metastatic disease and sociated with inoperability. If they are caused by more specifically with liver metastases. tumor, they probably are rising at a more rapid To determine operable cases one looks rate than those that occur later. Patients whose for patients with minimal elevations of CEA. In CEA elevations occur more than five months af­ one series the mean elevation in the operable ter surgery are more likely to be eligible for re­ patients was 6.5 ng/ml compared to 15.5 ng / exploration . These factors are summarized in ml in those who had inoperable malignancy.28 Table 6. Additional information can be gained by looking It is important to stress that the decision at the character of the rise . If a benign condition regarding reoperation is not based on just the exists, the CEA tends to plateau at levels usu­ CEA level. A careful determination must be ally less than 1 0 ng/ml. For any type of malig­ made that the patient does not have clin ical nant condition, an exponential rise occurs; the metastatic disease by obtaining a chest x-ray, slope of the rise for those who are operable is liver scan, serum chemistries , and other appro­ less than that for the inoperable ones. The value priate tests such as sonography, abdominal CT is probably in the range of 0.5 ng/ml/ mo or scan , and liver biopsy if other tests are nega­ less for operable cancers and >1 ng/ml/mo tive. Careful monitoring and testing will exclude for inoperable patients.29 five sixths of the patients for consideration for

TABLE 6 Differential Diagnosis of CEA Elevation

Time of Cause of CEA Liver Degree of Character Occurrence Elevation Nomogram Enzymes Elevation of Rise Post-Resection

Random Depends on Not Random Fluctuation - or± - Baseline Value Verifiable

Benign Often Usually Non Median Time Condition + + <1 Ong / ml Exponential <5mo. (Plateau) Operable + - Mean Exponential Malignancy 6.5 ng / ml Slope Median Time <1 ng / ml / mo >5mo. Inoperable Often mean Exponential Malignancy ++ + 15.5ng/ ml Slope ) > 1 ng / ml / mo

WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER / 95 "second-look" surgery. Of the remaining one nography and CT scanning are relatively new sixth, the resectability rate may be as high as procedures whose effect on the overall problem 30%. 30 Increasing the percentage of patients remains to be assessed . Carcinoembryonic an­ operated on will decrease the resectability per­ tigen testing is clearly an important advance, centage. The cure rate for the patients who are and there is every indication that other useful resected a second time is not known. markers will be developed. A practical strategy for following patients The net effect of all these developments is with colorectal cancer with CEA assays is pre­ the increased ability to select patients accu­ sented. Preoperative levels should be obtained rately for given treatments and to follow treat­ for establishing a base line and serve as a prog­ ment results more precisely. It is known from nostic indicator. The test should be repeated previous experience that in those diseases in postoperatively. Two weeks after surgery is which the assessment of results is difficult, both convenient and appropriate; however, if progress has been slow. Therefore, it is antici­ the CEA level has not returned to normal , the pated that more rapid improvements in treat­ test should be repeated. ments in the coming years will ultimately be re­ Apparently not all CEA values return to flected in the overall survival statistics of these normal promptly. If they remain elevated, it is diseases. important to establish a base line for nomogram analysis. Subsequently, values are obtained every two months for the early detection of re­ currence. After one year the test could be run REFERENCES less frequently. If a significant elevation is en­ countered outside the normal nomogram range , 1. Cancer Facts and Figures. American Cancer Society, the test is repeated serially, two or three times, 1980. to verify elevation and determine, if possible, the character of the rise. Concurrently, the pa­ 2. AXTELL LM , ASIRC AJ, MYERS MH: Cancer Patient Sur­ vival Report Number 5. DHEW Publication no. (NIH) tient is evaluated carefully for metastatic dis­ 77-992. National Cancer Institute, Bethesda, MD, ease. Selected patients may be candidates for 1976. surgery. In patients who have developed meta­ 3. MCDONALD JS, SCHEIN PS , WOOLLEY PV, ET AL: Five­ fluorouracil (5-FU), Mitomycin-C (MMC) and Adriamy­ static disease, CEA can be used for evaluation cin (ADR) FAM Combination Chemotherapy Results in of chemotherapeutic or radiotherapeutic re­ 61 Patients with Advanced Gastric Cancer. Pro­ sponse. The CEA appears to be more corre­ ceedings of the American Society of Clinical Oncology lated with tumor burden if its value is below 1 00 20:396, 1979. ng/ml. While this correlation is sometimes er­ 4. SCHAUER P, MAGILL GB, HOWARD J, ET AL: Combination ratic, it is the most valuable tool available in pa­ Chemotherapy of Gastric CA with MIFA II or with tients who do not have measurable disease. AAFC-CPPD . Proceedings of the American Society of Clinical Oncology 20:335, 1979. CONCLUSION Chemotherapy has produced significant 5. MOERTEL CG , O'CONNELL MJ, LAVIN PT: Chemotherapy of Gastric Cancer. Proceedings of the American Asso­ improvement in treatment results for gastric car­ ciation for Cancer Research 20:288, 1979. cinoma, but to date only minimal improvement has been achieved for colorectal carcinoma. 6. MOERT EL CG: Current concepts in cancer chemother­ Earlier application of radiation therapy, specifi­ apy of gastrointestinal cancer. N Engl J Med cally preoperative and postoperative radiation 229:1049-1952, 1978. therapy, particularly for patients with carcinoma 7. HAHN RG , MOERTEL CG , SCHUTT AJ , ET AL: A double­ of the rectum, is sufficiently attractive for further blind comparison of intensive course 5-fluorouracil by study. oral versus intravenous route in the treatment of colo­ The primary area of improvement for pa­ rectal carcinoma. Cancer35:1031-1035, 1975. tients with colorectal carcinoma has been in our 8. MOERTEL CG , SCHUTT AJ , HAHN RG , REITMEIER RJ: ability to assess the status of the disease and in Therapy of advanced colorectal cancer with a combi­ our beginning understanding of the biologic dif­ nation of 5-fluorouracil , methyl-1 , 3-cis (2-chloroethyl)- ferences in patients with the disease. Ultraso- 1 -nitrosourea, and vincristine. JNCI 54: 69- 71 , 1 9 7 5.

96 / WAMPLER: RECENT ADVANCES IN GASTROINTESTINAL CANCER 9. FALKSON G, FALKSON HC: Flu6rouracil, methyl-CCNU, 21. MACKAY IA : Use of Carcinoembryonic Antigen in and vincristine jn cancer of the colon. Cancer Screening an Unselected Population: A Five Year Fol­ 38:1468~1470, 1976. lowup in Clinical Application of Carcinoembryonic Anti­ gen Assay. Proceedings of a Symposium held in Nic;;e , 1 0. MACDONAl,D JS, KISNER OF, SMYTHE T, ET AL : Five-fluo­ France, Oct. 7-9, 1977, vol 439, pp 419-421 , Am­ rouracil (5-FU), methyl-CCNU and vincristine in the sterdam, Excerpta Medica International Congress Se­ treatment of advanced colorectal cancer: Phase II ries, 1978. study utilizing weekly 5-FU. Cancer Treat Rep 60:1597-1600, 1976. 22. LOEWENSTEIN MS , ZAMCH!=CK N: Carcinoembryonic an­ tigen (CEA) levels in benign gastrointestinal disease 11 . MOERTEL CG: Clinical management of advanced gas­ states. Cancer42(3):1412-1418, 1978. trointestinal cancer. Cancer 36:675-682, 1975.

23. ZAMCHECK N: The present status of CEA in diagnosis, 1 2. GREENE MH , BOICE JD, KEEHN RJ , ET AL: Late Effects of prognosis, and evaluation of therapy. Canc er Low Dose Adjuvant Chemotherapy in Colorectal Can­ 36:2460-2468, 1975. cer. Proceedings of the American Society of Clinical Oncology 20:413, 1979. 24. RAMMING KP , MACINTYRE J, ZAMCHECK N, ET AL: Serum 13. LI MC, Ross ST: Chemoprophylaxis for patients with carcinoembryonic antigen (CEA) monitoring of patients colorectal cancer: Prospective study with five-year fol­ at high risk for recurrence following surgery for colo­ low-up. JAMA 234:2825-2828, 1976. rectal carcinoma. Proceedings of the American So­ ciety of Clinical Qncology 20:329, 1 979. 14. MAVLIGIT GM, BURGESS MA, SEIBERT GB, ET AL: Prolon­ gation of postoperative disease-free interval ahd sur­ 25. GITNICH GL, MOLNAR IG: Carcinoembryonic antigen: vival in human colorectal cancer by B.C.G. or B.C.G. Trirnsmission by blood products. Cancer 42(3): 1 568- plus 5-fluorouracil. Lancet 1 :1248, 1976. 1573, 1978.

15. LAWRENCE W JR , TERZ JJ, HORSLEY S Ill: Chemotherapy 26. MARTIN EW, JAMES KJ , HURTUBISE PE, ET AL: The use of as an adjuvant to surgery for colorectal cancer. Ann CEA as an early indicator of gastrointestinal tumor re­ Surg 181 :616-623, 1975 currence and second-look procedures. Cancer 39:440-446, 1977. 1 6. HIGGINS GA JR , DWIGHT AW, SMITH JV, ET AL : Fluoroura­ cil as an adjuvant to surgery in carcinoma of the colon . 27. RI TTGERS RA, STEELE G JR , ZAMCHECK N, ET AL: Tran­ Arch Surg 102:339-343, 1971 . . sient carcinoembryonic antigen (CEA) elevations fol­ lowing resection of colorectal cancer: A limitation in 1 7. HIGGINS GA JR, HUMPHREY E, JULER GL, ET AL: Adjuvant the use of serial CEA levels as an indicator for second­ chemotherapy in the surgical treatment of colorectal look surgery JNCI 61 :315-318, 1978. cancer. Cancer38:1461-1467, 1976.

18. BLIKHINA NG, GARIN AM, LIPATOU AM: Results of Five 28. MINTON JP, MARTIN EW JR : The use of serial CEA de­ Year Observation for Patients Receiving 5-Fluorouracil terminations to predict recurrence of colon cancer and After Radical Surgery for Carcinoma of the Colon and when to do a second-look operation. Cancer Rectum. Proceedings of the Second All-Union Cancer 42:1422-142 7, 1978. Chemotherapy Conference, Kiev, September 1974, pp 243-244. 29. STAAB HJ , ANDE VER A, STUMPF E, ET AL: Slope analysis of the postoperative CEA time course and its possible 1 9. GRAGE TB, METTER GE, CORNELL GN , ET AL: The role of application as an aid in diagnosis of disease progres­ 5-fluoroi.Jracil as an adjuvant to the surgical treatment sion in gastrointestinal cancer. Am J Surg 136:322- of large bowel cancer. Adjuvant Therapy of Cancer, 327, 1978. Salmon SE, Jones SE (eds). Amsterdam, E;lsevier / North Holland, pp 259-263, 1977. 30. WILSON RE , PERENCEVICH NP, OLSON R, ET AL: Colorec­ tal adenocarcinoma: Patterns of metastases after cura­ 20. WEISS RB , DEVITA VT: Multimodal primary cancer treat­ tive resection and the role of serial CEA measurements ment (adjuvant chemotherapy): Current results and fu­ in their management. Eur Surg Res 1 0: 11 5-11 6 , ture prospects. Ann Intern Med 91 :251-256, 1979. 1978.

WAMPLER: RECENT ADVANCES IN GASTROINTE;STINAL CANCER / 97