computed tomography in the preoperative evaluation of patients with uterine corpus cancer

Available online at www.sciencedirect.com

Gynecologic Oncology 108 (2008) 486 – 492 www.elsevier.com/locate/ygyno

Comparison of the validity of magnetic resonance imaging and positron emission tomography/computed tomography in the preoperative evaluation of patients with uterine corpus cancer Jeong-Yeol Park a , Eyu Nyong Kim b , Dae-Yeon Kim a,⁎, Dae-Shik Suh a , Jong-Hyeok Kim a , Yong-Man Kim a , Young-Tak Kim a , Joo-Hyun Nam a a

Obstetrics and Gynecology, College of Medicine, University of Ulsan, Asan Medical Center, #388-1 Poongnap-dong, Songpa-Ku, Seoul, 138-736, South Korea b Department of Nuclear Medicine, University of Ulsan College of Medicine, Asan Medical Center, Seoul, South Korea Received 11 October 2007 Available online 16 January 2008

Abstract Objective. To compare positron emission tomography/computed tomography (PET/CT) with magnetic resonance imaging (MRI) in the preoperative detection of primary lesions and lymph node (LN) and distant metastases in patients with uterine corpus cancer. Methods. The patient cohort consisted of 53 women with uterine corpus cancer who underwent preoperative workup, including both MRI and PET/CT scans, and underwent surgical staging, including pelvic and/or paraaortic LN dissection, between October 2004 and June 2007 at Asan Medical Center, Seoul, Korea. Pathologic data from surgical staging were compared with the preoperative MRI and PET/CT results. For area specific analysis, LNs were divided into paraaortic, right pelvic and left pelvic areas. Results. In detecting primary lesions, MRI and PET/CT showed no differences in sensitivity (91.5% vs. 89.4%), specificity (33.3% vs. 50.5%), accuracy (84.9% vs. 84.9%), positive predictive value (PPV) (91.5% vs. 93.3%) and negative predictive value (NPV) (33.3% vs. 37.5%). With MRI, the sensitivity, specificity, accuracy, PPV and NPV for detecting metastatic LNs on LN area-by-area analysis were 46.2%, 87.9%, 83.9%, 28.6% and 94.0%, respectively; With PET/CT, those were 69.2%, 90.3%, 88.3%, 42.9%, and 96.6%, respectively. PET/CT showed higher sensitivity, but it did not reach statistical significance (p = 0.250). There were also no differences in specificity, accuracy, PPV and NPV. In detecting distant metastasis, the sensitivity, specificity, accuracy, PPV and NPV of PET/CT were 100%, 93.8%, 92.5%, 62.5% and 100%, respectively. Conclusion. In patients with uterine corpus cancer, PET/CT had moderate sensitivity, specificity and accuracy in detecting primary lesions and LN metastases, indicating that this method cannot replace surgical staging. The primary benefit of PET/CT is its sensitivity in detecting distant metastases. Because of its high NPV in predicting LN metastasis, PET/CT may also have advantages in selected patients who are poor candidates for surgical staging. © 2007 Elsevier Inc. All rights reserved. Keywords: Uterine corpus cancer; PET/CT; MRI; Preoperative evaluation; Lymph node

Introduction Uterine corpus cancer is a common malignancy of the female genital tract, which is staged using the surgical staging system of the International Federation of Obstetrics and Gynecology (FIGO) [1]. The routine use and extent of lymphadenectomy has not yet been established in all patients with uterine corpus

⁎ Corresponding author. Fax: +82 2 476 7331. E-mail address: [email protected] (D.-Y. Kim). 0090-8258/$ - see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ygyno.2007.11.044

cancer. Among the methods used to preoperatively assess the presence of pelvic and paraaortic lymph nodes (LNs) are imaging modalities, such as magnetic resonance imaging (MRI) and computed tomography (CT), but the results of these methods are not satisfactory [2,3]. Positron emission tomography (PET), which measures the increased metabolic activity of malignant cells rather than anatomic alterations, has been expected to have higher sensitivity than MRI or CT. In cervical, ovarian and vulvar cancer, many studies defining the validity of PET in diagnosing and staging patients have been published [4]. However, the role of PET in uterine corpus cancer is less

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492

defined because of the lack of published data in the literature [4]. To overcome the lower spatial resolution of PET compared with MRI or CT, a more advanced technique, fused positron emission tomography/computed tomography (PET/CT) was introduced [5,6]. This method combines the anatomic detail provided by CT with PET metabolic information [5,6]. To our knowledge, however, there have been no studies of PET/CT in the preoperative evaluation of patients with uterine corpus cancer. We therefore assessed the validity of PET/CT in the preoperative evaluation of patients with uterine corpus cancer and we compared the validity parameters of PET/CT with those of MRI. Materials and methods Study population This clinicopathologic comparative study involved 53 patients with uterine corpus cancer who underwent a preoperative workup that included both MRI and PET/CT scans and who underwent surgical staging that included pelvic and/ or paraaortic lymph node dissection between October 2004 and June 2007 at Asan Medical Center, Seoul, Korea. Their medical records were retrospectively reviewed, and surgical staging results were compared with preoperative findings on MRI and PET/CT. Initially, MRI was performed for preoperative evaluation for patients with endometrial cancer in our center. And, PET/CT has been included in preoperative evaluation of these patients since the introduction of PET/CT in our center. Form then on, MRI and PET/CT were recommended to most of patients who have uterine corpus cancer for preoperative evaluation. Therefore, MRI and PET/CT were performed in these patients except for patients who refuse this recommendation.

Preoperative evaluations All patients were diagnosed as having uterine corpus cancer by assessment of an endometrial biopsy, obtained using dilatation of the cervix and curettage of the endometrium. MRI was performed using a Signa 1.5-T scanner (General Electric Medical Systems). T1-weighted spin-echo images were obtained from the renal hila to the femoral neck using a body coil (TR range/TE range, 400–600/8–11; section thickness, 7–8 mm; intersection gap, 1 mm; field of view, 32 cm; number of acquisitions, 2; matrix, 256 × 256). Transverse and sagittal T2-weighted fast spin-echo images were obtained from the aortic bifurcation to the femoral neck using a body coil (4000–5000/90–110; echo-train length, 8; section thickness, 5–6 mm; intersection gap, 1 mm; field of view, 32 cm; number of acquisitions, 3; matrix, 512 × 512). Contrast-enhanced MR images were also obtained. Combined PET/CT was performed using a Biograph Sensation 16 scanner (Siemens/CTI, Knoxville, TN, U.S.). Patients were required to fast for 6 h prior to scanning, and combined PET/CT scanning from the skull base to the upper thighs was performed approximately 1 h after intravenous injection of 555 MBq FDG. PET/CT imaging was obtained by reconstruction of helical CT scans and PET images. The PET/CT scanner provides an in-plane spatial resolution of 6.0 mm, an axial field view of 16.2 cm, and three-dimensional image acquisition in six or seven bed positions with 2 min of emission. Whole-body CT scanning was performed in spiral mode with 100 mAs, 120 kVp, a section width of 5 mm, 0.75 mm collimation, and a table feed of 15 mm/gantry rotation, immediately preceding the acquisition of PET emission data. PET images were reconstructed using CT data for attenuation correction with the OSEM algorithm (2 iterations, 14 subsets) and 6.0 mm Gaussian filter using a 128 × 128 matrix. All PET/CT images were interpreted by the same experienced nuclear medicine physician who was aware of the patient's clinical history and results of previous imaging studies. All MRI images were interpreted by two experienced radiologists who were also aware of the patient's clinical history and results of previous imaging studies.

487

Surgical staging and pathologic evaluation All patients underwent surgical staging after preoperative evaluation. Standard procedures for surgical staging of uterine corpus cancer consisted of peritoneal washing cytology, exploration of peritoneal cavity, biopsy of any suspicious areas, simple hysterectomy, bilateral salpingo-oophorectomy and bilateral pelvic and/or paraaortic LN dissection through a laparotomic or laparoscopic route. When indicated, radical hysterectomy, omentectomy or cytoreduction of intra-abdominal disease was performed. Hysterectomy specimens were submitted for intraoperative frozen section examination. Bilateral pelvic LN dissection was performed from the deep circumflex iliac vessel to the mid-portion of the common iliac vessel. All LNs around the external, internal and common iliac vessels and in the obturator fossa were removed. Bilateral paraaortic LN dissection was performed from the midportion of the common iliac vessel to the renal vein. All LNs around the aorta and inferior vena cava and the aortocaval area between both ureters were removed. LNs were labeled as right and left external iliac, internal iliac, obturator, common iliac and paraaortic and divided into three areas (right pelvic, left pelvic and paraaortic) for evaluation. Each primary tumor and LN was sliced and stained with hematoxylin and eosin, and examined microscopically by a pathologist. The numbers of LNs retrieved in each area and the presence or absence of metastases were recorded.

Statistical analysis Surgicopathologic findings constituted the reference standard for statistical analysis. Comparisons were performed on LN area-by-area as well as patientby-patient bases. Differences in sensitivity, specificity and accuracy between MRI and PET/CT were estimated using the McNemar exact test. P-values (from two-sided tests) less than 0.05 were considered statistically significant. Data were analyzed using SPSS software for Windows (version 10.0; SPSS Inc., Chicago, IL).

Results The patient population consisted of 53 women with uterine corpus cancer, who were diagnosed by endometrial biopsy and underwent preoperative evaluation, including both MRI and PET/CT, between endometrial biopsy and staging operation. Median patient age was 52 years (range, 27–68 years) and median body mass index was 24 kg/m2 (range, 18–40 kg/m2). Ten patients were nulliparous. The median time from endometrial biopsy to both MRI and PET/CT was 8 days (range, 1– 27 days), and the median time from MRI and PET/CT to staging operation was 10 days (range, 1–28 days) and 9 days (range, 1– 23 days), respectively. International Federation of Obstetrics and Gynecology (FIGO) stage and pathologic characteristics are listed in Table 1. All 53 patients underwent pelvic LN dissection and 31 underwent paraaortic LN dissection. Final pathology reports showed no residual tumor in the hysterectomy specimens from six patients. A total of 1464 LNs were retrieved from 137 LN areas. The mean number of LNs retrieved per patient was 30 (range, 14–74). A total of 63 metastatic LNs were identified in 8 of 53 patients (18%) and in 13 of the 137 LN areas (12%). Validity parameters of MRI and PET/CT in detecting primary lesions Following endometrial biopsy, six patients had falsenegative results in detecting residual primary lesions in the

488

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492

Table 1 Patients' characteristics (n = 53)

of 8.1, and was not correlated with FIGO stage, histologic type or grade of tumor.

Characteristics Age (years) BMI (kg/m2) Parity (n)

FIGO stage (n)

Histology (n)

Grade of tumor (n)

Tumor size (cm) Myometrial invasion (n)

Values Median (range) Median (range) 0 1 2 ≥3 I II III IV Endometrioid Serous Mixed epithelial a MMMT ESS Leiomyosarcoma 1 2 3 NA Mean (range) No invasion Superficial Deep

52 (27–68) 24 (18–40) 10 4 18 21 31 7 8 7 32 8 1 8 2 2 14 12 19 8 4.2 (0.2–10.5) 10 23 20

TAH LAVH RH LRH BSO USO PLND PALND Omentectomy

28 19 5 1 45 2 53 31 17

Procedures performed (n)

BMI, body mass index; FIGO, International Federation of Obstetrics and Gynecology; MMMT, malignant mixed müllerian tumor; ESS, endometrial stromal sarcoma; NA, not applicable; SD, standard deviation; TAH, total abdominal hysterectomy; LAVH, laparoscopic-assisted vaginal hysterectomy; RH, radical hysterectomy; BSO, bilateral salpingo-oophorectomy; USO, unilateral salpingooophorectomy; PLND, pelvic lymph node dissection; PALND, paraaortic lymph node dissection. a Serous type and clear cell type.

endometrial cavity on PET/CT and/or MRI; three of these patients had false-negative results on both PET/CT and MRI. Final pathology report revealed thin superficially-spreading tumors in two of these patients and a 2.0 cm sized focal residual tumor in one patient. Two patients had false-negative results on PET/CT only; these patients had 0.7 cm and 0.6 cm diameter focal residual tumors in their hysterectomy specimens. One patient had a false-negative result on MRI only; she had a 0.3 cm sized focal residual tumor in her hysterectomy specimen. As shown in Tables 2 and 3, the sensitivity, specificity, accuracy, positive predictive value (PPV) and negative predictive value (NPV) of MRI in detecting residual primary lesions after endometrial biopsy (91.5%, 33.3%, 84.9%, 91.5% and 33.3%, respectively) did not differ from those of PET/CT (89.4%, 50.5%, 84.9%, 93.3% and 37.5%, respectively). The maximal standardized uptake value (maxSUV) of the primary lesions in PET/CT ranged from 1.4 to 21.1, with a mean

Validity parameters of MRI and PET/CT in evaluating pelvic and paraaortic LNs MRI detected 1 of 7 paraaortic LN areas (14.3% sensitivity) identified as metastatic according to pathologic examination, whereas PET/CT detected 4 of these 7 LN areas (57.1% sensitivity). Although PET/CT was more sensitive than MRI in predicting metastatic paraaortic LN, the difference was not statistically significant (p = 0.250). MRI resulted in 2 falsepositive interpretations (91.7% specificity), whereas PET/CT resulted in 3 false-positives (87.5% specificity) (p = 1.000). The accuracy, PPV and NPV of MRI vs. PET/CT in predicting paraaortic LN metastases were 72% vs. 76% (p = 0.625), 33.3% vs. 50% and 77.3% vs. 84.2%, respectively (Tables 2 and 3). Both MRI and PET/CT detected 5 of 6 pelvic LN areas (83.3% sensitivity) identified as metastatic according to pathologic examination (p = 1.000). MRI resulted in 13 false-positive interpretations (87.9% specificity), whereas PET/CT resulted in 9 false-positives (90.3% specificity) (p = 0.125). The accuracy, PPV and NPV of MRI vs. PET/CT in predicting pelvic LN metastasis were 80.8% vs. 88.5% (p = 0.125), 35.7% vs. 38.5% and 98.4% vs. 98.5%, respectively (Tables 2 and 3). For total LN areas, the sensitivity, specificity, accuracy, PPV and NPV of MRI vs. PET/CT in predicting LN metastasis were 46.2% vs. 69.2% (p = 0.250), 87.9% vs. 90.3% (p = 0.289), 83.9% vs. 88.3% (p = 0.065), 28.6% vs. 42.9% and 94.0% vs. 96.6%, respectively (Tables 2 and 3). In detecting LN metastasis using MRI, validity parameters for sarcoma and carcinoma were as follows: sensitivity, 50% vs. 55%; specificity, 75% vs. 84%; PPV, 50% vs. 22%; NPV, 75% vs. 94%; accuracy, 67% vs. 78%. In detecting LN metastasis using PET/CT, the validity parameters for sarcoma and carcinoma were as follows: sensitivity, 50% vs. 75%; specificity, 75% vs. 81%; PPV, 67% vs. 38%; NPV, 78% vs. 91%; accuracy, 75% vs. 85%. In a patient-by-patient comparison, the sensitivity, specificity, accuracy, PPV and NPV of MRI and PET/CT in predicting LN metastasis were 50.0% vs. 62.5% (p = 1.000), 80.0% vs. 86.7% (p = 0.375), 75.5% vs. 83.0% (p = 0.219), 30.8% vs. 45.5%, and 90.0% vs. 92.9%, respectively (Tables 2 and 3). For PET/CT, the maxSUV of metastatic LNs confirmed by pathology ranged from 2.1 to 15.2 (mean, 6.3), and the maxSUV of false-positive LNs ranged from 1.5 to 6.4 (mean, 3.1). Validity parameters of PET/CT in evaluating extrauterine disease Fourteen foci in 8 patients showed increased uptake of 2[18F]fluoro-2-deoxy-D-glucose (FDG), in the liver (n = 3), lung (n = 3), bone (n = 1), peritoneum, mesentery and omentum (n = 3), descending colon (n = 1), ovary (n = 1), thymus (n = 1), supraclavicular and mediastinal LN (n = 1) and axillary LN node (n = 1) (Fig. 1). Further evaluation of these lesions, by biopsy, colonoscopy, bone scan and subsequent imaging, showed that these lesions represented liver metastasis in 3 patients (max-

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492

489

Table 2 Comparison of the results of MRI, PET/CT, and pathology in detecting primary lesion, pelvic and praraaortic lymph node metastasis MRI

For primary lesion (n = 53)

Pathology

For PALN areas (n = 31)

Pathology

For PLN areas (n = 106)

Pathology

For total LN areas (n = 137)

Pathology

In patient-by-patient analysis (n = 53)

Pathology

Positive Negative Total Positive Negative Total Positive Negative Total Positive Negative Total Positive Negative Total

PET/CT

Positive

Negative

Total

Positive

Negative

Total

43 4 47 1 2 3 5 13 18 6 15 21 4 9 13

4 2 6 6 22 28 1 87 88 7 109 116 4 36 40

47 6 53 7 24 31 6 100 106 13 124 137 8 45 53

42 3 45 4 3 7 5 9 14 9 12 21 5 6 11

5 3 8 3 21 24 1 91 92 4 112 116 3 39 42

47 6 53 7 24 31 6 100 106 13 124 137 8 45 53

MRI, magnetic resonance imaging; PET/CT, positron emission tomography/computed tomography, PALN, paraaortic lymph node; PLN, pelvic lymph node; LN, lymph node.

SUV = 22.1, 7.7 and 4.1), lung metastasis in 2 patients (maxSUV = 6.8 and not reported), 1 patient with a benign lung lesion (maxSUV = 1.1), bone metastasis (maxSUV = 11.9), peritoneal metastasis (maxSUV = 7.7), mesenteric metastasis (maxSUV = Table 3 Comparison of validity parameters between MRI and PET/CT in evaluating primary lesion and lymph node status Validity parameters For primary lesion (n = 53)

For PALN areas (n = 31)

For PLN areas (n = 106)

For total LN areas (n = 137)

On patient-by-patient analysis a (n = 53)

Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV

MRI (%)

PET/CT (%)

P-value

91.5 33.3 84.9 91.5 33.3 14.3 91.7 74.2 33.3 78.6 83.3 87.0 87.7 27.8 98.9 46.2 87.9 83.9 28.6 94.0 50.0 80.0 75.5 30.8 90.0

89.4 50.5 84.9 93.3 37.5 57.1 87.5 80.6 57.1 87.5 83.3 91.0 88.1 35.7 98.9 69.2 90.3 88.3 42.9 96.6 62.5 86.7 83.0 45.5 92.9

1.000 1.000 1.000

0.250 1.000 0.625

1.000 0.125 0.125

0.250 0.289 0.065

1.000 0.375 0.219

MRI, magnetic resonance imaging; PET/CT, positron emission tomography/ computed tomography; PALN, paraaortic lymph node; PPV, positive predictive value; NPV, negative predictive value; PLN, pelvic lymph node; LN, lymph node. a In detecting metastatic LNs.

3.5), colon adenoma (maxSUV = 9.8), ovarian metastasis (maxSUV = 7.7), thymoma (maxSUV = 2.3), metastatic supraclavicular and mediastinal LN (maxSUV = 6.4), and reactive hyperplasia of axillary LN (maxSUV = 2.5). On a patient-by-patient basis, the sensitivity, specificity, accuracy, PPV, and NPV of PET/CT in evaluating extrauterine disease were 100%, 93.8%, 92.5%, 62.5% and 100%, respectively. Discussion PET is a molecular imaging technique that uses radiolabeled molecules to image molecular interactions of biologic processes. Among several PET compounds available, the radiolabeled glucose analog FDG is the only tracer approved by the US Food and Drug Administration (FDA) for routine clinical use [7]. Tumor cells are associated with increased glycolysis and this metabolic property leads to increased uptake of radiolabeled FDG [8]. Although enhanced FDG uptake is not specific to malignant cells, FDG-PET is widely used in many malignant diseases [4], because it is considered better than conventional CT or MRI in detecting metastatic or recurrent malignancies. Currently, the clinical use of FDG-PET has been approved in breast cancer patients, and it is under investigation in other malignancies [7,9], including gynecologic malignancies such as cervical, ovarian and vulvar cancer [4]. Like other malignancies, uterine corpus cancer also shows increased rates of glycolysis [10] and FDG uptake [11–15]. To date, however, only 2 retrospective and 1 prospective case series have investigated the role of PET in post-therapy surveillance of patients with uterine corpus cancer [11,12,16], and 2 prospective and 1 retrospective study have evaluated the sensitivity and specificity of PDF-PET in the preoperative detection of pelvic and paraaortic LN metastases [16–18]. Since PET has a limited ability to provide information on the exact location of a lesion, fused PET/CT, which combines the anatomic detail provided by CT with PET metabolic informa-

490

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492

Fig. 1. Preoperative combined PET/CT showing advanced endometrial cancer with multiple metastases. (A) PET image (coronal view). (B) combined PET/CT image. (C) CT image. (D) PET image (maximal intensity projection image). (a) supraclavicular lymph node metastasis (maxSUV = 6.4), (b) liver metastasis (maxSUV = 4.2), (c) paraaortic lymph node metastasis (maxSUV = 15.2), (d) primary lesion in the endometrial cavity (maxSUV = 16.8) and pelvic lymph node metastasis (maxSUV = 4.5).

tion, has been developed as an alternative to CT or MRI. PET/ CT also has the benefit of attenuation correction [4–6]. However, only one retrospective case series has investigated the role of PET/CT in the post-therapy surveillance of patients with uterine corpus cancer [19], and there have been no studies in-

vestigating the sensitivity and specificity of preoperative PET/ CT in these patients. To our knowledge, this retrospective analysis is the first to compare the sensitivity, specificity, accuracy, PPV and NPV of PET/CT with those of MRI in the preoperative evaluation of patients with uterine corpus cancer.

Table 4 Reports on validity parameters of PET or PET/CT in preoperative evaluation for primary uterine corpus cancer Author (year) [reference]

Number of patients

Imaging modality

Validity parameters (%)

Primary lesion (%)

PALN (%)

PLN (%)

Total LN (%)

Extrauterine disease

Horowitz et al. (2004) [17]

19 a

PET

Chao et al. (2005) [16]

27 b

PET

Suzuki et al. (2007) [18]

30 c

PET

Present study

53

PET/CT

Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV Sensitivity Specificity Accuracy PPV NPV

84 – – – – – – – – – 96.7 – – – – 89.4 50.5 84.9 93.3 37.5

– – – – – 85 95 93 – – 0 100 94.7 0 94.7 57.1 87.5 80.6 57.1 87.5

– – – – – 73 95 90 – – 0 100 80.8 0 80.8 83.3 91.0 88.1 35.7 98.9

60 98 – – – – – – – – – – – – – 69.2 90.3 88.3 42.9 96.6

– – – – – – – – – – 83.3 100 – 100 96.0 100 93.8 92.5 62.5 100

PALN, paraaortic lymph node; PLN, pelvic lymph node; LN, lymph node; PET, positron emission tomography; PET/CT, positron emission tomography/computed tomography; PPV, positive predictive value; NPV, negative predictive value. a One patient did not undergo staging operation due to distant metastasis. b Four patients did not undergo lymph node dissection. c The number of patients who underwent lymph node dissection in primary staging operation was not reported.

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492

In evaluating primary lesions in the endometrial cavity, we found that PET/CT and MRI showed similar sensitivity, but that neither was able to detect very small focal lesions and thin superficially-spreading tumors. The limited sensitivity (84%) of PET alone in detecting primary small sized lesions of the uterus with minimal FDG uptake [17] was not increased in PET/CT. We found that the mean maxSUV of uterine corpus cancer in the endometrial cavity was 8.1 (range, 1.4–21.1), in good agreement with previous reports (mean, 8.8 (range, 7–13.5) [14], mean, 7.5 (range, 3.8–16.8) [15], mean, 4.5 (range, 3.0–6.3) [13] and mean, 7.8 (range, 2.7–17.4) [18]). In comparison, the mean maxSUVs were lower in patients with benign endometrial lesion (4.5, range 2.3–8.2) and normal menstrual cycles (menstrual flow phase: mean 5, range 2.3–16.6; proliferative phase: mean 2.6, range 1.1–5.7; ovulatory phase: mean 3.7, range 2– 5.4; secretory phase: mean 2.5, range 1.3–5.6). PET/CT showed higher sensitivity than MRI in detecting paraaortic LN metastases, although the difference did not reach statistical significance (57.1% vs. 14.3%, p = 0.250). In comparison, one series found that the sensitivity of PET in detecting paraaortic LN metastases was higher than that of MRI or CT (85% vs. 46%) [16], whereas another reported that the sensitivity of PET was much lower than that of CT (0% vs. 100%) [18]. In detecting pelvic LN metastasis, we found that PET/CT and MRI showed the same sensitivity (83.3%); in comparison, one series found that the sensitivity of PET was greater than that of MRI or CT (73% vs. 53%) [16], whereas another found that the sensitivity of PET was lower than that of MRI or CT (0% vs. 40%) [18]. In total, we found that PET/CT had a higher sensitivity than MRI in detecting paraaortic or pelvic LN metastasis, although the difference did not reach statistical significance (69.2% vs. 46.2%, p = 0.250). In a prospective study including 19 patients with primary uterine corpus cancer, PET had moderate sensitivity in detecting paraaortic or pelvic LN metastases [17]. Using these previous assessments of PET alone for comparison, we found that PET/CT did not show added benefits in predicting LN status (Table 4). Moreover, PET/CT had only slightly higher sensitivity than MRI in detecting LN metastasis, suggesting that PET/CT cannot replace surgical staging of LN status. However, PET/CT had high specificity and NPV in detecting LN metastasis, suggesting it may allow the safe omission of lymphadenectomy in selected patients who are poor candidates for surgical staging [17]. Interestingly, we found that both MRI and PET/CT showed very low PPV in predicting LN metastases, perhaps because all patients underwent endometrial biopsy using dilatation and curettage prior to imaging, with the resultant reactive lymphadenopathy misinterpreted as LN metastasis. We found that PET/CT detected all true metastatic lesions, but that the false-positive rate was somewhat high. Unlike other imaging techniques, PET/CT accurately localized the sites of extraperitoneal metastatic tumors over the entire body, sites that were not detected by conventional imaging studies that cover only the abdomen and pelvis. Moreover, PET/CT can detect asymptomatic metastatic lesions at the time of preoperative evaluation. Our findings suggest that the main benefit of PET/

491

CT, like that of PET alone [16,18], is its ability to sensitively detect, localize and characterize distant metastases. To overcome limited sensitivity and specificity of imaging technique using FDG, other PET compounds such as O-(2-[18F] fluoroethyl)-L-tyrosine (18F-FLT), 11C-acetate, 11C-choline, [18 F]fluoromisonidazole and CU-labeled diacetyl-bis(N(4)mehtylthiosemicarbazone may be useful in gynecologic cancer [4]. In addition, recent development of nanoparticulate contrast agents in conjunction with MRI may be helpful [20]. Further investigations with these novel imaging techniques and agents should be conducted. One of the principal limitations of our study was its retrospective design. In addition, the PET/CT and MRI images were interpreted by radiologists and a nuclear medicine physician who were aware of the patient's clinical history and the results of previous imaging modalities, thus possibly introducing some bias. However, MRI was performed first, and then PET/CT was performed. Therefore, radiologists did not know the results of PET/ CT when they interpret the MRI. In addition, the radiologist did not amend the result of MRI after PET/CT. One of the strengths of our study is its assessment of a large number of patients. In conclusion, PET/CT had moderate sensitivity, specificity and accuracy in preoperative detection of primary lesions and LN metastases in patients with uterine corpus cancer. Its sensitivity in detecting residual intrauterine malignancy after endometrial biopsy was similar to that of MRI, and its sensitivity in predicting LN metastasis was slightly higher than that of MRI, but the difference was not statistically significant. These findings indicate that PET/CT cannot replace surgical staging. The principal benefit of PET/CT is its sensitivity in detecting distant metastases. Due to its high NPV in predicting LN metastases, PET/CT can also be useful in selected patients who are poor candidates for surgical staging. The low PPV of this method may be due to misinterpreting reactive lymphadenopathy after endometrial biopsy as malignant change. Larger prospective studies are needed to clarify the role of PET/CT in preoperative evaluation of uterine corpus cancer. References [1] International Federation of Gynecology and Obstetecis. Corpus cancer staging. Int J Gynaecol Obstet 1989;28:190. [2] Connor JP, Andrews JI, Anderson B, Buller RE. Computed tomography in endometrial carcinoma. Obstet Gynecol 2000;95:692–6. [3] Hricak H, Rubinstein LV, Gherman GM, Karstaedt N. MR imaging evaluation of endometrial carcinoma: results of an NCI cooperative study. Radiology 1991;179:829–32. [4] Lai CH, Yen TC, Chang TC. Positron emission tomography imaging for gynecologic malignancy. Curr Opin Obstet Gynecol 2007;19:37–41. [5] Schoder H, Erdi YE, Larson SM, Yeung HW. PET/CT: a new imaging technology in nuclear medicine. Eur J Nucl Med Mol Imaging 2003;30: 1419–37. [6] Wahl RL. Why nearly all PET of abdominal and pelvic cancers will be performed as PET/CT. J Nucl Med 2004(45 Suppl 1):82S–95S. [7] Juweid ME, Cheson BD. Positron-emission tomography and assessment of cancer therapy. N Engl J Med 2006;354:496–507. [8] Wahl RL, Hutchins GD, Buchsbaum DJ, Liebert M, Grossman HB, Fisher S. 18F-2-deoxy-2-fluoro-D-glucose uptake into human tumor xenografts. Feasibility studies for cancer imaging with positron-emission tomography. Cancer 1991;67:1544–50. [9] Cachin F, Prince HM, Hogg A, Ware RE, Hicks RJ. Powerful prognostic

492

[10] [11]

[12]

[13]

[14]

[15]

J.-Y. Park et al. / Gynecologic Oncology 108 (2008) 486–492 stratification by [18F]fluorodeoxyglucose positron emission tomography in patients with metastatic breast cancer treated with high-dose chemotherapy. J Clin Oncol 2006;24:3026–31. Warburg O. The metabolism of tumors. London: Arnold Constable; 1930. p. 75–327. Belhocine T, De Barsy C, Hustinx R, Willems-Foidart J. Usefulness of (18) F-FDG PET in the post-therapy surveillance of endometrial carcinoma. Eur J Nucl Med Mol Imaging 2002;29:1132–9. Saga T, Higashi T, Ishimori T, Mamede M, Nakamoto Y, Mukai T, et al. Clinical value of FDG-PET in the follow up of post-operative patients with endometrial cancer. Ann Nucl Med 2003;17:197–203. Umesaki N, Tanaka T, Miyama M, Kawamura N, Ogita S, Kawabe J, et al. Positron emission tomography with (18)F-fluorodeoxyglucose of uterine sarcoma: a comparison with magnetic resonance imaging and power Doppler imaging. Gynecol Oncol 2001;80:372–7. Lerman H, Metser U, Grisaru D, Fishman A, Lievshitz G, Even-Sapir E. Normal and abnormal 18F-FDG endometrial and ovarian uptake in preand postmenopausal patients: assessment by PET/CT. J Nucl Med 2004; 45:266–71. Torizuka T, Kanno T, Futatsubashi M, Okada H, Yoshikawa E, Nakamura

[16]

[17]

[18]

[19]

[20]

F, et al. Imaging of gynecologic tumors: comparison of (11)C-choline PET with (18)F-FDG PET. J Nucl Med 2003;44:1051–6. Chao A, Chang TC, Ng KK, Hsueh S, Huang HJ, Chou HH, et al. 18FFDG PET in the management of endometrial cancer. Eur J Nucl Med Mol Imaging 2006;33:36–44. Horowitz NS, Dehdashti F, Herzog TJ, Rader JS, Powell MA, Gibb RK, et al. Prospective evaluation of FDG-PET for detecting pelvic and paraaortic lymph node metastasis in uterine corpus cancer. Gynecol Oncol 2004;95:546–51. Suzuki R, Miyagi E, Takahashi N, et al. Validity of positron emission tomography using fluoro-2-deoxyglucose for the preoperative evaluation of endometrial cancer. Int J Gynecol Cancer Mar 2007;2 [Electronic publication ahead of print]. Sironi S, Picchio M, Landoni C, Galimberti S, Signorelli M, Bettinardi V, et al. Post-therapy surveillance of patients with uterine cancers: value of integrated FDG PET/CT in the detection of recurrence. Eur J Nucl Med Mol Imaging 2007;34:472–9. Mulder WJ, Griffioen AW, Strijkers GJ, Cormode DP, Nicolay K, Fayad ZA. Magnetic and fluorescent nanoparticles for multimodality imaging. Nanomed 2007;3:307–24.