ANTICANCER RESEARCH 27: 557-562 (2007)

Scintimammography with a Hybrid SPECT/CT Imaging System

ORAZIO SCHILLACI, ROBERTA DANIELI, LUCA FILIPPI, PASQUALE ROMANO, ELSA COSSU, CARLO MANNI and GIOVANNI SIMONETTI

Department of Biopathology and Diagnostic Imaging, University "Tor Vergata", Rome, Italy

Abstract. Background: Planar scintimammography is useful carcinomas are evident on mammograms, especially in dense for characterizing lesions >10 mm. Our aim was to or dysplastic (3); moreover, its specificity and positive evaluate Tc-99m sestamibi scintimammography with a hybrid predictive value are low because it cannot always (SPECT/CT) device for functional anatomical mapping (FAM). differentiate benign lesions from malignant ones (4). The Patients and Methods: Three planar images and a chest drawbacks of have led to the development of SPECT/CT were performed with a hybrid device in 53 patients complementary modalities for breast cancer imaging, with mammographically suspicious lesions. The final including scintimammography, a nuclear technique histopathological diagnosis was obtained after surgery. Results: that uses to detect malignant breast The planar images were positive in 27 out of 37 carcinomas tumours (5, 6). Scintimammography is conventionally (sensitivity 73%) and the SPECT/CT in 33 (sensitivity 89.2%). performed with planar acquisitions, which have a low The sensitivity of planar imaging and SPECT/CT was 42.9% sensitivity for lesions ≤10 mm (5). Recently, a new imaging and 71.4% in cancers ≤10 mm, and 91.3%, and 100% in device combining a dual-head, variable angle cancers >10 mm, respectively. The specificity was 93.8% for with a low-dose X-ray tube has been introduced (7): this both planar and SPECT/CT imaging; accuracy was 79.2% for hybrid gamma camera/CT scanner provides cross-sectional planar scans and 90.6% for SPECT/CT. FAM was useful in X-ray transmission images that facilitate anatomical providing a precise anatomical localisation of the SPECT localisation of radiotracer uptake, because the acquired findings. Conclusion: SPECT/CT scintimammography using a tomoscintigraphic and CT images are co-registered by means hybrid device is able to detect breast cancer, showing a sensitivity of the hardware in the same session (8). higher than that of planar images, especially for small cancers. The aim of this study was to evaluate the usefulness of this new imaging system for functional anatomical mapping Breast cancer is the most common malignancy in women: it (FAM) in patients with suspected breast cancer submitted represents 32% of all invasive tumours in the female to scintimammography with Tc-99m sestamibi, and if FAM population in the U.S.A., and it accounts for 15% of all is able to improve the accuracy of conventional planar women’s cancer deaths. In 2005, the American Cancer Society images in breast cancer diagnosis. estimated 211,240 new cases and 40,410 female deaths from this disease (1). Trends in incidence and mortality show that Patients and Methods there has been a small but steady annual increase in breast carcinoma incidence over the last 30 years, whereas the Fifty-three female patients (age range: 27-78 years) with suspicious mortality rate has declined steadily since the beginning of 1990 breast lesions on mammography were evaluated. The final (1). This improvement is attributed to earlier detection of histopathological diagnosis was obtained in all the cases after surgery. breast cancer using mammographic screening. Each patient gave written informed consent to participate in the study. Patients were injected with 740 MBq of Tc-99m sestamibi Till now, mammography is the best imaging modality for ~ (Bristol-Myers Squibb Pharma, Bruxelles, Belgium) into an the early identification of breast cancer (2). Nevertheless, antecubital vein of the opposite arm to the known mammary lesion. this technique has some limitations: not all breast The labelling and quality control procedures were carried out according to the manufacturer's instructions. All scintigraphic images were acquired using a combined SPECT/CT system (Millenium VG & Hawkeye; General Electric Medical Systems, Correspondence to: Prof. Orazio Schillaci, Viale G. Mazzini 121, Milwaukee, WI, USA) composed of a dual-head, variable angle 00195 Rome, Italy. Tel: +39 06 20902419, Fax: +39 06 37359664, gamma-camera and an X-ray tube with a set of detectors mounted e-mail: [email protected] on opposite sides of the gamma-camera gantry. This system enables the sequential interchangeable acquisition of and Key Words: Breast cancer, scintimammography, Tc-99m sestamibi, CT images. The CT scanner has a fixed anode oil-cooled X-ray hybrid imaging, SPECT/CT. tube operating at 140 kV and 2.5 mA during the acquisition of each

0250-7005/2007 $2.00+.40 557 ANTICANCER RESEARCH 27: 557-562 (2007)

Table I. Histopathological findings. Table II. Overview of the results.

Histology Lesion (n) Planar FAM

Malignant lesions 37 Sensitivity 73% (27/37) 89.2% (33/37)* Infiltrating ductal carcinoma 29 Mixed ductal and lobular carcinoma 4 Specificity 93.8% (15/16) 93.8% (15/16) Infiltrating lobular carcinoma 3 PPV 96.4% (27/28) 97.1% (33/34) Infiltrating mucinous carcinoma 1 NPV 60% (15/25) 78.9% (15/19) Benign conditions 16 Accuracy 79.2% (42/53 90.6% (48/53)* Fibrocystic disease 8 Fibroadenoma 6 FAM=functional anatomical mapping; PPV=positive predictive value; Papilloma 1 NPV=negative predictive value. Cyst 1 *p<0.05 vs. planar.

transverse slice. The X-ray tube and the linear detector array rotate significance of the differences between the sensitivity, specificity together in a fixed geometry at 2.6 revolutions per minute, with the and accuracy of SPECT/CT and planar scintimammography for the acquisition of a single slice in 14 seconds. After completion of the detection of breast cancer. The statistical significance of the first acquisition (transmission or emission), the patient is differences between the sensitivity of the two procedures was automatically repositioned so that the previous 40-cm axial field calculated taking into account overall primary breast cancers and exactly corresponds with the second imaging modality’s 40-cm axial after subdividing the carcinomas according to palpability (palpable field of view. vs. non-palpable) and size (≤10 mm vs. >10 mm). The ¯2 test was The scintigraphic protocol included multiple planar views, used to assess the statistical significance of differences in the starting 10 minutes post-injection, followed by SPECT/CT imaging. sensitivity of SPECT/CT and planar images after subdividing the Three planar scans (128x128 matrix, 10 min/view) were performed: breast cancers according to palpability and size. P-values below two lateral images (left and right) with the patient in the prone 0.05 were considered significant. position with a special cushion that enables the examined breast to hang freely through an opening, very close to the collimator of the Results gamma-camera, and an anterior chest image with the patient in the supine position. SPECT/CT was acquired with the patients in the supine position using these parameters: SPECT over 360Æ (matrix 128x128, 3Æ According to the histopathologic results, the patients were angle steps, 30 sec/frame), followed by CT, obtained by "half-scan" divided into two groups as malignant and benign breast lesions acquisition over 220Æ for each transaxial slice; multiple slices were (Table I). There were 37 cancers (four different histologic performed by moving the table by a slice step before acquiring the types, palpable in 21 cases and non-palpable in 16 cases) and next slice, with the full field of view consisting of 40 slices. 16 benign lesions (four different histologic diagnoses, palpable From the raw data, SPECT images were reconstructed to in 11 cases and non-palpable in 5 cases). According to the produce transaxial slices; coronal and sagittal views were then American Joint Committee on Cancer (AJCC) tumour (T) obtained. Also CT data were reconstructed using the nuclear medicine workstation (eNTEGRA, General Electric Medical pathologic classification, three carcinomas were classified as Systems, Milwaukee, WI, USA) to produce cross-sectional T1a, 11 as T1b, 18 as T1c and five as T2. attenuation images (256x256 matrix) in which each pixel represents SPECT/CT scintimammography detected 33 out of 37 the attenuation of the imaged tissue. The X-ray reconstructed primary breast carcinomas, planar imaging 27/37, and both images were implemented into nuclear medicine. Matching pairs procedures were true-negative in 15/16 benign mammary of emission and transmission were fused and images overlying the lesions. Table II reports the overall sensitivity, specificity, SPECT data on the corresponding anatomic planes were accuracy, positive and negative predictive values of generated. All were evaluated blindly by two experienced SPECT/CT and planar scintimammography, calculated on nuclear medicine physicians (O.S. and R.D.) who did not know the the basis of true-positive, false-negative, true-negative and final diagnosis of the patients, and classified as positive or negative false positive results. for Tc-99m sestamibi focal abnormal accumulation in the breast; SPECT/CT was positive in 1/3 T1a, in 9/11 T1b, in 18/18 disagreements were resolved by consensus, with a third observer T1c and in 5/5 T2 carcinomas, while planar images were (L.F.) as a referee. positive in 0/3, in 6/11, in 16/18 and in 5/5, respectively. When considering palpability, SPECT/CT in all the palpable Statistical analysis. The scintigraphic results were compared to the histopathological findings after surgery, and the sensitivity, cancers and in 12 out of 16 of the non-palpable ones, specificity, accuracy, and positive and negative predictive value whereas planar imaging was positive in 19/21 and in 8/16, were determined. McNemar’s test was used to assess the statistical respectively. SPECT/CT sensitivity values were significantly

558 Schillaci et al: SPECT/CT Scintimammography higher than those of the planar scans both in cancers sized of a North American multi-centre trial involving 673 <10 mm (71.4% vs. 42.9%, p<0.05) and in non-palpable patients in 42 institutions (11) have indicated a sensitivity ones (75% vs. 50%, p<0.05). Moreover, in all the malignant for breast cancer detection of 87% and 61% for palpable tumours positive both at SPECT7CT and planar imaging, and non-palpable abnormalities, respectively. SPECT/CT gave a better anatomical localization of the Technical considerations for scintimammography should be lesions. Figures 1 and 2 illustrate two patients with two T1b taken into account to obtain good results; in particular, the carcinomas, both positive at SPECT/CT and negative at acquisition of tomographic images, by means of SPECT, could planar imaging. play a role in increasing the sensitivity of planar images. An SPECT/CT scintimammography was false-negative in experimental study carried out on a breast phantom four breast cancers: two T1a non-palpable invasive ductal demonstrated a statistically significant gain in the detection carcinomas and two T1b non-palpable invasive carcinomas rate of small size lesions using SPECT imaging when (one ductal and one lobular). Planar images were false- compared to planar scans (12). However, in clinical practice, negative in all the above mentioned four cases missed by SPECT acquisition is not routinely performed, and the value SPECT/CT and also in 6 further malignant lesions: 1 T1a of this imaging method has not yet been definitively non-palpable invasive ductal carcinoma, three T1b invasive established (5). Therefore, the potential role of SPECT in ductal carcinomas (two non-palpable and one palpable), two increasing the sensitivity of planar scintimammography, T1c cancers (one non-palpable invasive ductal carcinoma especially for small-sized tumours, is still controversial. In and one palpable invasive lobular carcinoma). comparative studies using SPECT for primary breast cancer The only false-positive case observed with both SPECT/CT detection, discordant findings have been reported: SPECT and planar imaging was a palpable fibroadenoma (size 14 sensitivity has been found to be lower than planar images by mm), highly suspicious for malignancy at mammography, some authors (13-15) and higher by other investigators (16- which showed moderate signs of inflammation and a high 18). To this end, it is important to highlight that good quality proportion of proliferating cells at histopathology. SPECT imaging can be obtained only with the patient in the supine position and the arms up, because SPECT with Discussion patients in the prone position is limited by geometric constraints of the patient, the imaging table and the gantry (5). The capacity of visualizing small carcinomas is critical for the The main drawback of SPECT scintimammography is the clinical development and acceptance of scintimammography, absence of precise anatomical landmarks that renders because the other modalities are increasingly localisation efforts difficult and often inaccurate (19). In used for the early identification of small suspicious fact, although SPECT provides better contrast resolution abnormalities. Mammography is currently the imaging than planar images, it can be difficult to obtain accurate modality of choice in detecting early, non-palpable breast localisation of breast lesions. Fusion imaging can solve this cancer: however, false-negative results are reported, as well problem by combining the functional information of nuclear as false-positive findings, and scintimammography has medicine studies with the anatomical details of CT or MRI, proved to be a very useful adjunct to a non-diagnostic or so allowing a precise localisation of sites of increased difficult mammography (5). uptake: anatomic landmarks are Neverthless, the literature data clearly indicate that the important in order to precisely identify the sites of abnormal sensitivity of conventional planar scintimammography is uptake and the structures containing normal activity. These strictly dependent on the size of lesions. A multi-centre are the basis for the growing interest in the development of study on 420 patients with 449 breast lesions (355 multi-modality devices combining structural and functional carcinomas) has reported sensitivity of 26%, 56%, 95% and images. These hybrid systems are able to acquire data in the 97% for T1a, T1b, T1c and T2 breast cancers, respectively same scanning session, and therefore to reduce the (9); in particular, sensitivity was significantly different problems often observed with software approaches to image between malignant lesions >10 mm (96%) and those ones fusion (20). In particular, the first clinical reports indicate sized ≤10 mm (46.5%). Similar findings regarding that the dual-modality data acquired using a device sensitivity are obtained when breast lesions are grouped in combining a dual-head, variable angle gamma camera with palpable and non-palpable, which always show a lower a low-dose X-ray tube, is clinically useful in cancer imaging sensitivity on breast , as confirmed by Liberman with single-photon radiopharmaceuticals, including Tc-99m et al. (10), who, in a recent meta-analysis considering 64 sestamibi (21, 22). unique studies with data on 5,340 patients with 5,354 breast The present study was performed with this hybrid lesions (malignant n=3024), have reported sensitivity of SPECT/CT imaging system in a series of patients with 87.8% for patients with a palpable breast mass and of 66.8% suspected breast cancer. The data obtained have clearly for patients without a palpable lesion. Moreover, the results indicated the clinical usefulness of Tc-99m sestamibi

559 ANTICANCER RESEARCH 27: 557-562 (2007)

Figure 1. Sixty-one years old patient with a proven non-palpable 6 mm infiltrating ductal carcinoma in the external quadrants of the right breast. Mammography in the mediolateral oblique projection (A) demonstrates a nodular spiculated opacity (black arrow). Tc-99m sestamibi planar scintimammography (right prone lateral view) is negative (B). SPECT (C, center) shows a focal uptake (white arrow) that SPECT/CT (C, right) demonstrates corresponding to the carcinoma (white arrow); the respective CT scan is on the left.

SPECT/CT scintimammography. The comparative evaluation These results suggest that the use of SPET/CT must be of FAM and planar acquisitions have demonstrated the preferred for small non-palpable breast cancers, because its greater effectiveness of SPECT/CT. In particular, a marked sensitivity is significantly higher than that of planar scans. difference in sensitivity both in non-palpable primary breast Moreover, also in malignant lesions of >10 mm the planar cancers and in those sized ≤10 mm, the two categories that imaging detection rate was slightly lower than that of FAM. are most difficult to detect and to distinguish from benign In our series, both SPECT/CT and planar lesions with conventional diagnostic procedures was found. scintimammography demonstrated high specificity (93.8% Planar images missed more than 50% of these lesions, and for both) and positive predictive value (97.1% and 96.4%, they were never positive in carcinomas negative on FAM. respectively), with the same single false-positive result in

560 Schillaci et al: SPECT/CT Scintimammography

Figure 2. Fifty-six years old patient with a proven non-palpable 10 mm infiltrating ductal carcinoma in the lower quadrants of the left breast. Mammography in the craniocaudal projection (A) demonstrates a nodule with irregular borders (black arrow). Tc-99m sestamimibi planar imaging (left prone lateral view) is negative (B), while SPECT (C, center) shows a focal area of increased uptake uptake (white arrow) that SPECT/CT (C, right) precisely localizes in the breast (white arrow); the corresponding CT scan is on the left.

one palpable high-proliferating fibroadenoma with signs inflammatory component of benign lesions playing a of inflammation. This outcome is in accordance with the major role (23). findings of a recent paper aimed of assessing the tissue- In conclusion, this study confirms the high specificity of specific effects on uptake of Tc-99m sestamibi by breast Tc-99m scintimammography in differentiating malignant lesions, with a targeted analysis of false scintigraphic from benign breast lesions in patients with suspected breast diagnoses; overall, the uptake pattern was determined by carcinoma. Moreover, our findings indicate that SPECT/CT a combination of factors, with the extent of the with a hybrid device is useful for breast imaging: it improves

561 ANTICANCER RESEARCH 27: 557-562 (2007) the diagnostic accuracy of conventional planar 12 Garin E, Devillers A, Girault S et al: Scintimammography: scintimammography. In particular, FAM is more sensitive better detection of small-sized lesions with tomoscintigraphic than planar scintigraphy, especially in the detection of non- than planar images, a phantom study. Nucl Med Commun 22: 1045-1054, 2001. palpable and/or sized ≤10 mm breast cancers. Therefore, a 13 Palmedo H, Schomburg A, Grunwald F et al: Technetium-99m- more extensive use of this new technology for breast MIBI scintimammography for suspicious breast lesions. J Nucl imaging with scintimammography is suggested; however, we Med 37: 626-630, 1996. are aware that further studies in larger series are needed to 14 Tiling R, Tatsch K, Sommer H et al: Technetium-99m-sestamibi confirm our data. scintimammography for the detection of breast carcinoma: comparison between planar and SPECT imaging. J Nucl Med References 39: 849-856, 1998. 15 Danielsson R, Bone B, Agren B et al: Comparison of planar and SPECT scintimammography with 99mTc-sestamibi in the 1 Jemal A, Murray T, Ward E et al: Cancer statistics, 2005. CA diagnosis of breast carcinoma. Acta Radiol 40: 176-180, 1999. Cancer J Clin 55: 10-30, 2005. 16 Helbich TH, Becherer A, Trattnig S et al: Differentiation of 2 Tabar LK, Vitak B, Chen HHT et al: Beyond randomized benign and malignant breast lesions: MR imaging versus Tc-99m controlled trials: organized mammographic screening sestamibi scintimammography. 202: 421-429, 1997. substantially reduces breast cancer mortality. Cancer 91: 1724- 17 Aziz A, Hashmi R, Ogawa Y et al: Tc-99m-MIBI 1731, 2001. scintimammography; SPECT versus planar imaging. Cancer 3 Berg WA, Gutierrez L, NessAiver MS et al: Diagnostic accuracy Biother Radiopharm 14: 495-500, 1999. of mammography, clinical examination, US, and MR imaging 18 Spanu A, Dettori G, Nuvoli S et al: 99mTc-tetrofosmin SPET in preoperative assessment of breast cancer. Radiology 233: in the detection of both primary breast cancer and axillary 830-849, 2004. lymph node metastasis. Eur J Nucl Med 28: 1781-1794, 2001. 4 Kopans D: The positive predictive value of mammography. Am 19 Khalkhali I, Mena I, Diggles L: Review of imaging techniques J Roentgenol 158: 521-526, 1992. for the diagnosis of breast cancer: a new role of prone 5 Schillaci O and Buscombe JR: Breast scintigraphy today: scintimammography using technetium-99m sestamibi. Eur J indications and limitations. Eur J Nucl Med Mol Imaging 31 Suppl Nucl Med 21: 357-362, 1994. 1: S35-S45, 2004. 20 Schillaci O and Simonetti G: Fusion imaging in nuclear 6 Taillefer R: Clinical applications of 99mTc-sestamibi medicine – applications of dual-modality systems in . scintimammography. Semin Nucl Med 35: 100-115, 2005. Cancer Biother Radiopharm 19: 1-10, 2004. 7 Bocher M, Balan A, Krausz Y et al: Gamma camera-mounted 21 Schillaci O, Danieli R, Manni C et al: The added clinical value anatomical X-ray : technology, system characteristics of SPET/CT with a hybrid imaging system. J Nucl Med 44: and first images. Eur J Nucl Med 27: 619-627, 2000. 290P, 2003. 8 Schillaci O: Hybrid SPECT/CT: a new era for SPECT imaging? 22 Schillaci O, Danieli R, Manni C et al: Is SPECT/CT with a Eur J Nucl Med Mol Imaging 32: 521-524, 2005. hybrid camera useful to improve scintigraphic imaging 9 Scopinaro F, Schillaci O, Ussof W et al: A three center sudy on interpretation? Nucl Med Commun 25: 705-710, 2004. the diagnostic accuracy of Tc-99m MIBI scintimammography. 23 Tiling R, Stephan K, Sommer H et al: Tissue-specific effects on Anticancer Res 17: 1631-1634, 1997. uptake of 99mTc-sestamibi by breast lesions: a targeted analysis 10 Liberman M, Sampalis F, Mulder DS et al: Breast cancer of false scintigraphic diagnoses. J Nucl Med 45: 1822-1828, 2004. diagnosis by scintimammography: a meta-analysis and review of the literature. Breast Cancer Res Treat 80: 115-126, 2003. 11 Khalkhali I, Villanueva-Meyer J, Edell SL et al: Diagnostic accuracy of 99mTc-sestamibi breast imaging: multicenter trial Received July 27, 2006 results. J Nucl Med 41: 1973-1979, 2000. Accepted September 18, 2006

562