Interference of Known or Suspected Endometriosis in Reporting FDG PET/CT Performed in Another Indication

We performed a retrospective search in our prospective databases including electronic reports of 18 F-FDG PET/CTs, using the keyword “endometriosis,” appearing in the summary of patient's history and/or as a potential diagnosis matching the images. This search was performed in 2 centers and covered the years 2010 to 2020 and retrieved the 18 F-FDG PET/CT scans of 21 patients. The criteria for inclusion in the study were that endometriosis was part of the past or present history of the patient or was highly suspicious at the time of 18 F-FDG PET/CT, which confirmed later that endometriosis was not the (main) indication for 18 F-FDG PET/CT and that follow-up data were available. Eighteen patients (9 from each center) met those criteria. This search overall resulted in a series of 18 patients with known or suspected endometriosis subsequently confirmed referred to 18 F-FDG PET/CT for search and characterization of lesions in the context of cancer staging or follow-up, active granulomatous disease, or inflammatory syndrome (Table 1 ). Patient's Demographics, Clinical Context, and Result of 18 F-FDG PET/CT *Known. †Suspected. The indication for 18 F-FDG PET/CT, the age of the patient, and the delay since the diagnosis of endometriosis were recorded. 18 F-FDG PET/CT was performed during the proliferative phase of the cycle in 15 of 16 patients and during the menstrual flow in 1 of 16 nonovariectomized premenopausal women without oral contraception. Fifteen PET/CT procedures were performed on a Gemini TF-16 machine and 3 on a Siemens Biograph mCT flow machine, starting approximately 60 minutes after intravenous injection of 2 to 3 MBq/kg of body mass of 18 F-FDG. The field of view of PET covered the top of the skull to midthighs at 3-minute acquisition per bed position. The location and SUVmax of each nonphysiologic 18 F-FDG foci were recorded. For each patient, a composite standard of truth based on clinical follow-up, results of other imaging modalities, and biopsy or postsurgical histology was subsequently determined, in order to characterize the nature of those 18 F-FDG foci.

The patients' demographics, clinical context, and the result of 18 F-FDG PET/CT (lesions detected on 18 F-FDG PET/CT and their SUVmax of according to their nature) are provided in Table 1 . The data of 18 women with mean age at 18 F-FDG PET/CT of 42.5 years (median, 40 years; range, 26–66 years) were analyzed. Fifteen of them (83%) (patients 1–15) had a past or recent history of endometriosis. The mean time interval since the diagnosis of endometriosis was 5.4 years (median, 2 years; longest time interval of 40 years at the time of 18 F-FDG PET/CT); 4 of 18 (22%) (patients 1, 8, 12, and 13) of these patients had undergone unilateral or bilateral salpingo-oophorectomy or partial oophorectomy because of endometriosis. Endometriosis was suspected as one potential diagnosis in 3 of 18 patients (17%) (patients 16–18) referred to 18 F-FDG PET/CT for characterization of a pelvic lesion or staging of a pelvic tumor and was subsequently confirmed during follow-up. According to the composite standard of truth, the presence of lesions of endometriosis at the time of 18 F-FDG PET/CT was confirmed in 13 of 18 patients (72%) (patients 2–4, 6, 7, 9, 10, 12, 13, and 15–18). In contrast, in 1 patient, the left adnexal mass suspected to correspond to a late recurrence of her endometriosis was in fact an ovarian cancer with extension to the uterus. In 4 patients (5, 8, 11, and 14), no recurrence of endometriosis was detected either on other imaging modalities at the time of 18 F-FDG PET/CT or during follow-up. The per-patient positivity rate of 18 F-FDG PET/CT showing foci evocative of metabolically active lesions of endometriosis was 8 of 18 (44%; 95% confidence interval, 22–69%) (patients 2, 3, 7, 9, 10, and 16–18). The 18 F-FDG positivity of endometriosis was influenced neither by the menopausal nor oophorectomy status of the patient (Fisher test P = 0.3) nor by the time interval since the diagnosis of endometriosis (mean, 6.3 years if 18 F-FDG–negative vs. 4.3 years if 18 F-FDG–positive P = 0.5). The oldest patient with 18 F-FDG–positive endometriosis was aged 63 years; the longest time interval between the diagnosis of endometriosis and the detection of 18 F-FDG–positive foci of endometriosis was 15 years. The presence of pelvic implants of endometrium was detected in 4 patients on MRI (patients 4, 12, 13, and 15) and in 1 patient (patient 6) during surgery, but in those patients, no foci were visible on 18 F-FDG PET/CT. Thus, the patient-based detection rate of 18 F-FDG PET/CT limited to patients with confirmed lesions of endometriosis was 8 of 13 (62%; confidence interval, 32%–86%). On per-lesion/site basis, 18 F-FDG PET/CT detected 11 of 20 sites (55%) of endometriosis. The SUVmax of these lesions/sites ranged between 1.8 and 5.3 (median, 3.8). In 9 of 18 patients (50%) (patients 1, 3, 4, 8, 10, 11, 13, and 5), a total of 13 non–endometriosis-related lesions/sites were detected by 18 F-FDG PET/CT; their SUVmax ranged between 2.7 and 23 (median, 9.4). All 7 malignant lesions/sites were 18 F-FDG–positive (ovarian cancer, urothelial cancer, sarcomatous lung cancer with lymph node metastases, lymph node recurrence of breast cancer); their SUVmax ranged between 4.6 and 23 (median, 13). Three benign tumors took up 18 F-FDG: 2 myomas (SUVmax 4.4 and 7.0) and 1 breast fibroadenoma (SUVmax 2.7). Three inflammatory lesions also took up 18 F-FDG: colitis, sarcoidosis, and pleural inflammation after talc pleurodesis (SUVmax 6.2, 7.2, and 9.4, respectively). In 1 patient subsequently diagnosed with polychondritis, no 18 F-FDG focus was visible (Fig. 1 ). SUVmax of lesions detected on 18 F-FDG PET/CT according to their location and origin. The illustrative cases of 18 F-FDG–avid endometriosis in a patient with active (patient 3) and remittent malignancy (patient 7) are provided in Figure 2 and Figure 3 . The illustrative case of mildly 18 F-FDG–avid ovarian endometrioma in a patient with suspected ovarian cancer is provided in Figure 4 . 18 F-FDG PET/CT, A : MIP, B and E : PET, C and F : CT, and D and G : PET/CT, axial slice. Staging of sarcomatous lung cancer in a 47-year-old woman (patient 3) with 11-year history of endometriosis at the time of 18 F-FDG PET/CT. Intense 18 F-FDG uptake by primary lung cancer (SUVmax 23) and its lymph node and pleural metastases. Two foci of increased 18 F-FDG uptake in lesions of endometriosis in the uterine wall (SUVmax 4.3 and 5.3, B , D , and E , G , arrow). No 18 F-FDG uptake by known right ovarian endometrioma ( G , arrowhead). 18 F-FDG PET/CT, A : MIP, B : PET, C : CT, and D : PET/CT, axial slice. Surveillance of left breast cancer resected 3 years earlier in a 63-year-old woman (patient 7); letrozole since then. The endometriosis was known for 4 years at the time of 18 F-FDG PET/CT. Focally increased 18 F-FDG uptake in rectosigmoid junction (SUVmax 3.8; B and D , arrow) confirmed by biopsy as endometriosis. 18 F-FDG PET/CT, A : MIP, B : PET, C: CT, and D : PET/CT, axial slice. Characterization and staging of newly diagnosed left ovarian mass complicated by constriction of left ureter and hydronephrosis in a 47-year-old woman (patient 17) with no history of endometriosis. Mildly increased, isolated 18 F-FDG uptake in the peripheral part of cystoid left ovarian lesion (SUVmax 4.5, B and D : arrow) and increased 18 F-FDG uptake in uterine cavity during menstrual flow (SUVmax 7.45, B and D : arrowhead). Left ovarian endometrioma was confirmed by histology.

According to our results derived from 18 patients, the interference of known or suspected endometriosis in reporting 18 F-FDG PET/CT performed in another indication seems to be limited, but possible and should be kept in mind. Our results confirm that the lesions of endometriosis show inconstant 18 F-FDG uptake, which, in the majority of cases, is lower than that of malignant lesions, with partial overlap of SUVmax with well-differentiated, low-grade malignancies. According to the history of the patient, endometriosis should be considered as one potential origin of 18 F-FDG foci, even in postmenopausal women, as the oldest patient with 18 F-FDG–positive endometriosis was aged 63 years. In case endometriosis has been diagnosed several years ago, incidental 18 F-FDG–positive lesion may still correspond to endometriosis, the longest time interval being 15 years in our series. In our series, the greatest SUVmax value of lesion of endometriosis was 5.3, which is somewhat higher than the SUVmax threshold of 4 previously proposed for identification of MTOE.

In endometriosis, fibrosis and inflammation are likely to be the nonspecific diagnostic targets on FDG PET/CT, with a variable intensity of uptake (Table 2 ). Cases have been reported of 18 F-FDG uptake by progressive deep invasive pelvic endometriosis 11 or by endometriosis lesions in the lung, 12 the canal of Nuck, 20 or lymph nodes, 14 but in general, a significantly increased glucose metabolism does not seem to be a part of typical and consistent features of endometrioma. 21 , 22 In the study by Rieber et al, 8 18 F-FDG PET, performed for characterization of asymptomatic ovarian mass suspicious for carcinoma, showed positive foci in 4 of 22 patients (18%) with ovarian endometrioma. On the same year, the same team 10 reported increased 18 F-FDG uptake corresponding to the area of adnexal endometriosis in 5 of 23 cases (22%) only. Furthermore, when compared with MRI, the 18 F-FDG foci could be clearly attributed to endometrioma in 1 of 5 cases; in the remaining cases, the 18 F-FDG uptake was caused by gastrointestinal activity. In the subsequent study by Fastrez et al, 9 the preoperative 18 F-FDG PET/CT brought negative result in all 9 consecutive patients aged 31 ± 10 years with confirmed ovarian and/or pelvic endometriosis (6/9 had at least stage III endometriosis, according to the American Society of Reproductive Medicine). In the series by Fastrez et al 9 including patients with suspected endometriosis, 18 F-FDG uptake was observed in none of the lesions of the 9 patients with confirmed endometriosis. In the study by Setubal et al 13 including 9 patients with suspected endometriosis, 18 F-FDG foci were visible in 4 of 8 patients with confirmed endometriosis, allowing only a partial visualization of the disease spread in those 18 F-FDG–positive cases. 18 F-FDG Uptake by Lesions of Endometriosis *Endometriosis finally confirmed in a lesser number of patients. DR, detection rate; NA, not available. This inconstant 18 F-FDG uptake by lesions makes 18 F-FDG PET/CT not suited for routine localization of lesions of endometriosis. 3 , 4 In our series of patients referred to 18 F-FDG PET/CT for another indication but with past or recent diagnosis or suspected endometriosis, the patient-based detection rate of endometriosis with 18 F-FDG was 8 of 13 (62%), nonsignificantly greater than 32 of 73 (44%) obtained by pooling the results of 5 previous studies (Fisher exact test, P = 0.4). 8 – 10 , 13 , 18 It should be noted that the detection rate was highly variable between those series, from none of 9 (0%) 9 to 18 of 22 (82%) 8 in case of ovarian masses. On 18 F-FDG PET/CT requested in another indication, an incidental 18 F-FDG uptake by lesions of endometriosis, even if infrequent, 23 , 24 may be a source of misinterpretation. We will consider criteria that may help to circumvent this pitfall. Active endometriosis also occurs in 2% to 5% postmenopausal women 6 , 7 and commonly represents an adverse effect of hormone replacement therapy or tamoxifen treatment 25 – 27 ; in a few cases, postmenopausal endometriosis has been described in women who had no history of endometriosis on imaging or surgery prior to menopause. 26 Those results and a case report 17 justify considering endometriosis as a potential diagnosis and source of 18 F-FDG foci even in postmenopausal patients. In this cohort, we observed 18 F-FDG active lesions of endometriosis in 1 of 4 postmenopausal or postoophorectomy patients. Two postmenopausal patients with history of endometriosis were referred for localization of recurrent breast cancer. One patient was being treated with letrozole, an aromatase inhibitor that has an inhibitory effect on endometriosis, but that did not impede 18 F-FDG uptake in this patient. The other patient had been treated with tamoxifen more than 10 years before and no lesion of endometriosis was found. In a 66-year-old patient, the suspected ovarian endometriomas corresponded to an ovarian cancer infiltrating the uterus, and in a 63-year-old patient with confirmed endometriosis in uterosacral ligament, torus, and rectosigmoid junction, the 18 F-FDG PET/CT localized 1 lesion of endometriosis in the rectosigmoid junction. The common sites of endometriosis include ovaries, fallopian tubes, pelvic peritoneum, and pelvic ligaments, whereas atypical sites include the colon, urinary bladder, ureter, abdominal wall, and pleura. 28 , 29 For pelvic sites, endometriosis shares the typical localization with ovarian cancer, the abdominal cavity being the typical localization of metastases. 30 In the present series, the lesions of confirmed endometriosis were located not only in the pelvic peritoneum, adnexal area, and rectosigmoid junction, but also in the abdominal wall and in the thorax in 1 patient (patient 2). Non–endometriosis-related malignant, inflammatory, or benign 18 F-FDG–positive lesions were detected in 5 patients in the abdomen or the pelvis and also in 5 patients in the thorax (patients 2, 3, 8, 10, and 11). SUVmax values of endometriosis lesions reported in literature are summarized in Table 2 . In the published case reports, the SUVmax of lesions of endometriosis located in the pelvic or abdominal muscle ranged from 1.7 to 4.8. 11 , 15 , 16 In ovarian endometriosis, the SUVmax ranged from 1 to 4 in a series of 5 cases. 18 Among 8 patients with endometriosis, Setubal et al 13 observed 18 F-FDG uptake in 3 of 5 patients with deep endometriosis (SUVmax 4.88, 5.44, and 5.56). SUVmax of abdominal wall endometriosis was 3.52, whereas ovarian endometrioma and superficial endometriosis were misdiagnosed on 18 F-FDG PET/CT in 1 case each. A trend for a higher 18 F-FDG uptake by deep endometriosis, as measured by SUVmax, compared with superficial peritoneal lesions may be explained by a greater lesion size of deep endometriosis, whereas peritoneal lesions are often of small size. Another explanation is related to the type of peritoneal lesions red, black, or Wright stain–positive. Indeed, red lesions are very active lesions with a high angiogenesis, whereas Wright stain–positive lesions are mainly composed of fibrosis, and black lesions correspond to intermediate state, but also by chronic inflammatory and/or fibrotic character of the pathology. 3 , 23 In the present series, the SUVmax values of endometriosis lesions fell into the same range. The lesions of endometriosis with the highest 18 F-FDG uptake were located in pelvic peritoneum and adnexal area (SUVmax 2.3–5.3). In contrast, the lesions of endometriosis in the rectosigmoid junction (SUVmax 3.6–3.8), the abdominal wall (2.6), and the thorax (1.8–4.5) were less metabolically active. This finding may be explained by a larger fibrotic component and a lesser glandular component of endometriosis in those latter organs and structures. The SUVmax values reported in case of malignant transformation of endometriosis (MTOE) are provided in Table 3 . Malignant transformation of endometriosis is a particular clinical situation that occurs in fewer than 0.1% of patients. A significant difference in 18 F-FDG uptake ( P 4.0 was capable of excluding endometriosis, with 75% sensitivity, 100% specificity, 100% positive predictive value, and 68.8% negative predictive value (area under the curve 90%). 18 Accordingly, in a case of rectal MTOE reported by Li et al, 19 the SUVmax was 15.7. In another patient with MTOE in the abdominal wall, the SUVmax on 18 F-FDG PET/MRI was 9.61 in the abdominal wall and 4.25 in metastatic retroperitoneal and inguinal lymph nodes. 32 In the case reported by Yoshida et al, 31 18 F-FDG was taken up by MTOE in the groin, as well as by ipsilateral metastatic pelvic lymph nodes in a postmenopausal patient, but the 18 F-FDG uptake values were not provided. 18 F-FDG Uptake in Case of Malignant Transformation of Endometriosis Other types of malignancies usually result in a high value of SUVmax, but there is an overlap with some lesions of endometriosis. Just focusing on gynecological malignancies, Takagi et al 33 reported a mean SUVmax value of 11.0 (range, 2.6–22.4) in 41 ovarian cancers, 13.7 (2.0–35.2) in 63 endometrial cancers, and 15.9 (5.8–29.3) in 15 patients with uterine sarcomas. In this study, SUVmax values less than 5.6 were found only in ovarian cancers at International Federation of Gynecology and Obstetrics stage 1 (n = 19; mean, 9.8; range, 2.6–22.4) and stage 2 (n = 5; mean, 10.1; range, 4.0–18.6). Recently, Park et al 34 reported the SUVmax according to the dualistic model of ovarian carcinogenesis; SUVmax values less than 5.6 may be observed with type I (n = 90; mean, 9.5; range, 2.5–23.8) and type II (n = 80; mean, 13.0; range, 3.2–41.2). In the present series, the SUVmax of ovarian cancer lesions was 13 in the pelvis and 11.5 in the thorax; the SUVmax of sarcomatous lung cancer and its metastatic lymphadenopathy was 23 and 21.3, respectively. However, the SUVmax of a retroperitoneal low-grade urothelial tumor was 4.6, similar to the value observed by Maffione et al 16 in an infiltrative retroperitoneal endometriosis, overlapping with the range of SUVmax of endometriosis lesions. In the 3 patients who had 18 F-FDG–positive lesions originating from a malignancy and others from endometriosis, all the malignant lesions had a greater SUVmax than all the lesions of endometriosis. Apart from endometriosis, other benign gynecological tumors can take up 18 F-FDG. In the series of 47 patients with uterine myomas by Ma et al, 35 18 F-FDG positivity rate was 19/47 (40.4%). The SUVmax of 18 F-FDG–positive myomas was highly variable (4.38 ± 2.57; range, 1.4–10.81) and was higher in younger and premenopausal patients. 35 Takagi et al 33 reported 18 F-FDG uptake by 12 uterine myomas with SUVmax of 6.0 ± 5.4 (range, 2.7–22.4). Therefore, 18 F-FDG SUVmax of myomas seems to be similar to or higher than that of lesions of endometriosis. This inconstant and variable 18 F-FDG uptake by myomas may be explained by various levels of inflammation in the microenvironment suitable for myoma onset and development. 36 According to the results of the retrospective study by Nezhat et al 37 performed in 244 patients, 181 of 208 (87%) patients with a chief concern of symptomatic myomas also had histology-proven endometriosis. In the present study, 18 F-FDG–positive uterine myoma was found in 2 patients (patients 4 and 10), associated with lesions of endometriosis, which were 18 F-FDG–positive in 1 case. Therefore, in patients with uterine myoma, the presence of foci on 18 F-FDG PET/CT corresponding to lesions of endometriosis must be considered. 18 F-FDG PET provides a high accuracy in the differentiation of fibroadenomas from malignant tumors because fibroadenomas usually show no or mild 18 F-FDG uptake. 38 Accordingly, in our series, only mild 18 F-FDG uptake (SUVmax 2.7) was observed in 1 case of breast fibroadenoma. Concerning the differentiation between noninfected endometriosis and infectious or inflammatory lesions, the reported uptake values for endometriosis were lower than values reported in a patient with bilateral ovarian abscess (SUVmax 8.4), but similar to uptake values in reactive lymphadenopathy (SUVmax 2.6–5.7, mean SUVmax 4) in the same patient. 39 In the present series, the chronic inflammatory pleural reaction after talc pleurodesis, lesions of sarcoidosis in mediastinal lymph nodes, and nonspecific colitis showed higher SUVmax (9.4, 7.2, and 6.2, respectively) than any lesion of confirmed endometriosis. Finally, in a patient who was not treated with corticosteroids, evolutive polychondritis did not result in foci on 18 F-FDG PET/CT, in contrast with previously published results. 40