Introduction

Ovarian cancer is one of the most commonly diagnosed and lethal gynecological malignancies among females worldwide [1]. About half of the patients will relapse within 16 months after primary debulking surgery and platinum-based adjuvant chemotherapy [2]. Thus, effective indicators are urgently required to earlier estimate the biological behavior and treatment efficacy of tumor cells to guide individual treatments.

18F-FDG PET/CT is an important molecular imaging modality for preoperative and therapeutic effect assessment in ovarian cancer. Previous studies have shown that semi-quantitative parameters of PET/CT, such as SUVmax, SUVavg, and MTV, are associated with patients’ clinical characteristics and prognoses in ovarian cancer [3,4,5,6,7,8,9]. Several investigators have found that higher PET/CT parameters are associated with an advanced FIGO stage and a higher tumor burden and could indicate a poorer prognosis [7,8,9,10,11]. However, other studies have failed to draw similar conclusions [12, 13].

Moreover, tumor cell proliferation and chemosensitivity are also two vital factors of malignant tumors. Nevertheless, research on the relationship between PET/CT parameters and ovarian cancer’s proliferative activity and chemosensitivity is limited. Thus, our study aimed to explore both clinical and prognostic significance of 18F-FDG PET/CT in epithelial ovarian cancer.

Materials and methods

Clinicopathological data

We retrospectively investigated 48 epithelial ovarian cancer patients who had preoperative 18F-FDG PET/CT scan and subsequent primary debulking surgery between January 2010 and June 2015 in our hospital. All patients had PET/CT scan within 2 weeks before primary surgery. Patients were excluded if they received neoadjuvant chemotherapy, or were treated for recurrent disease. The study protocol was approved by the institution’s review board.

All pathological slides were independently reviewed by two experienced gynecologic pathologists according to the WHO criteria. Immunohistochemical staining was performed for Ki67 (Roche 30-9) using a Ventana Benchmark XT autostainer (Ventana Medical Systems Inc., Tucson, AZ, USA).

The clinical data were collected from medical records, cancer registries, and pathology reports. Surgical outcome has been classified as R0, R1 and R2, according to residual disease (RD). R0 indicated no macroscopic residual disease, while R1 and R2 were defined as the maximum RD diameter of < 1 cm or ≥ 1 cm, respectively. The optimal surgery comprised R0 and R1. Chemosensitivity was defined as that we previously reported, a time interval of 6 months or longer between the completion of platinum-based chemotherapy and recurrence. Otherwise, the patient was chemoresistant [14, 15]. All of the patients were followed up until December 31, 2016. Overall survival (OS) was defined as the time interval from the date of operation to the date of death or the last follow-up.

18F-FDG PET/CT protocol

18F-FDG was produced automatically by cyclotron (Siemens CTI RDS Eclipse ST, Knoxville, TN, USA) using the Explora FDG4 module in our center. Radiochemical purity was over 95%. All patients were asked to fast at least 6 h before the 18F-FDG PET/CT procedure. The patient’s blood glucose level was below 10 mmol/L at the time of radio-tracer injection. The standard dosage of intravenous 18F-FDG administration was 7.4 MBq/kg. Before and after injection, patients were kept lying comfortably in a quiet, dimly lit room. Examination was performed 60 min after 18F-FDG injection. A Siemens Biograph 16HR PET/CT scanner (Knoxville, TN, USA) with 4 mm transaxial intrinsic spatial resolution (full width at half maximum) and 16.2 cm axial field width was used for image scanning. Whole-body PET/CT data acquisition began with low-dose CT from the inguinal region to the head, with 120 kV, a 80–250 mA automatic adjustment, a pitch of 3.6, and a 0.5 s rotation time, followed by PET emission scan in a 3-dimensional mode, with 2–3 min per bed position. The PET data were reconstructed using the ordered-subset expectation maximization technique selecting eight subsets and four iterations, a 168 × 168 matrix. The CT data were used for attenuation correction of the PET images, and coregistered images were displayed on a workstation. The reconstructed images were then converted to a semiquantitative image corrected by the injection dose and the subject’s body weight (SUV).

Imaging interpretation

All PET/CT images were independently evaluated by two experienced nuclear medicine physicians. Consensus was reached in case of discrepancy. Analysis of the 18F-FDG PET/CT images was performed using a multimodality computer platform (Syngo, Siemens, Knoxville, TN, USA). SUVmax for each patient was calculated by placing a spheroid-shaped volume of interest (VOI) within the primary ovarian lesion. The SUV was calculated as [decay-corrected activity (kBq) per milliliter of tissue volume]/[injected 18F-FDG activity (kBq) per gram of body mass]. MTV and SUVavg were measured by drawing a contour of the ovarian lesion large enough to encase the primary tumor in the axial, coronal, and sagittal PET images. We used a threshold of SUV of 2.5 [16,17,18], which is usually used in clinical settings, to define the margins around the target. The boundaries of voxels, the SUV of which exceeded a threshold of 2.5 SUV, were automatically produced. The voxels presenting an SUV intensity of greater than 2.5 within the contouring margin were incorporated to define the MTV, and the mean voxels within this contouring were measured as SUVavg. The kidneys, ureters, and the urinary bladder were manually subtracted to ensure the correct voxel.

Statistical analysis

SPSS statistical software (version 21.0, SPSS, IBM Inc., Armonk, NY, USA) was used for statistical analyses. For demographic data, a descriptive statistical analysis method was used for analysis, and the results were summarized as means with standard deviations (SD), medians with ranges, or frequencies with percentages. The categorical data were compared using Chi-square or Fisher’s exact tests. Correlations between SUVmax, SUVavg, MTV, and Ki67 were tested using Pearson’s correlation analysis. OS was analyzed using the Kaplan–Meier method, and log-rank tests were used in the univariate analyses. P < 0.05 was considered statistically significant, and all P values reported were 2-sided.

Results

Patients’ characteristics

Patients’ characteristics are described in Table 1. Among the 48 patients, the median (range) age of the patients was 52 (27–76) years, and 43 (89.6%) had advanced FIGO (stage III–IV). The majority of the patients (93.8%) had high-grade serous ovarian cancer. The median (range) Ki67 index was 70% (15–95%).

Table 1 Patients’ characteristics
Full size table

Optimal surgery was achieved for 42 patients (87.5%), including 17 debulked to R0 and 25 to R1. All patients received platinum-based adjuvant chemotherapy after surgery. Among these, 6 patients had disease progression during chemotherapy, and 32 (66.7%) were sensitive to platinum-based chemotherapy. The median (range) follow-up time was 27 (8–61) months. Nineteen patients (39.6%) relapsed during the follow-up period. At the end of the follow-up, 30 patients were still alive, seven were dead, and the other 11 were censored.

PET/CT parameters and their associations with clinicopathological characteristics and prognosis

As for preoperative PET/CT scan parameters, the median (range) SUVmax, SUVavg, and MTV values were 11.42 (3.14–20.2), 4.8 (2.55–9.47), and 150.11 (0.19–792.46), respectively. We analyzed the relationships between the clinical features and the PET/CT parameters. SUVmax, SUVavg, and MTV had no association with age, FIGO stage, CA125 level, HE4 level, ascites and postoperative residual disease.

Overall, 83.3% of our patients had a high Ki67 index (> 50%), and the remaining seven patients had a moderate Ki67 index (5–50%). There was a significant positive correlation between the SUVmax value and the Ki67 index (P = 0.030, r = 0.314). Nevertheless, neither SUVavg nor MTV values were associated with Ki67 index (P = 0.051 and r = 0.283 and P = 0.586 and r = − 0.081, respectively) (Fig. 1). Furthermore, chemosensitive patients had higher SUVmax values than those who were chemoresistant. However, there were no differences in SUVavg and MTV values between these two groups (Table 2).

Fig. 1
figure 1

a–c Correlation between cell proliferation and SUVmax, SUVavg, and MTV. The SUVmax value had a positive correlation with the Ki67 index (a)

Full size image
Table 2 PET/CT parameters and chemosensitivity
Full size table

In addition, patients were dichotomized by the median of each PET/CT parameter into high and low level groups for prognostic analyses, respectively. There were no differences in OS between the high and low SUVmax, SUVavg, or MTV groups, respectively (Fig. 2).

Fig. 2
figure 2

a–c Kaplan–Meier curves of OS. None of these three PET/CT parameters were found to be potential predictors of OS

Full size image

Discussion

Our study demonstrated that the SUVmax value is associated with Ki67 index and chemosensitivity. However, neither SUVavg nor MTV had associations with patients’ clinicopathological parameters, and none of these three PET/CT parameters were found to be potential predictors of OS.

Compared to conventional imaging technologies such as CT, MR, and ultrasound, 18F-FDG PET/CT could not only assess anatomical or morphological features of malignant tumors, but also provide their functional or metabolic characteristics [19, 20]. PET/CT scan is used increasingly during ovarian cancer treatment for its noninvasive and multifunctional advantages [5, 10, 12, 21] (Figs. 3, 4).

Fig. 3
figure 3

a–c Representative transaxial CT (a), PET (b) and PET/CT fused (c) ovarian tumor (white arrow showed) images of a 68-year-old chemosensitive patient with a FIGO stage IIIC. Postoperative pathology was HGSC with a 85% Ki67 value. The patient underwent six cycles’ chemotherapy after debulking surgery, and had no recurrence and metastasis until the last follow-up date. The SUVmax, SUVavg and MTV of ovarian tumor were 19.2, 6.8 and 164.96, respectively

Full size image
Fig. 4
figure 4

a–c Representative transaxial CT (a), PET (b) and PET/CT fused (c) images of ovarian tumor (cross showed) in a 60-year-old chemoresistant patient staged IIIC. Postoperative pathology was HGSC with a 55% Ki67 value. The patient underwent eight cycles’ chemotherapy and recurred 12 months after cytoreductive surgery. The SUVmax, SUVavg and MTV were 6.4, 3.5 and 40.31, respectively

Full size image

The Ki67 index is a vital predictive and prognostic marker for malignant tumors [22]. It is routinely used to estimate tumor proliferative activity in daily clinical practice. Several studies have shown that PET/CT parameters are correlated with the Ki67 index and thus could indicate tumor proliferation in several types of cancer [22,23,24,25,26]. Data for ovarian cancer proliferation are limited. Our study found that SUVmax had a positive correlation with the Ki67 value, while SUVavg or MTV did not. This may be because higher proliferative tumor cells have increasing glucose transporter proteins and hexokinase and glucose-6-phosphatase activities [27]. This will lead to higher 18F-FDG uptake accumulation, which could be directly reflected by SUVmax. However, SUVavg and MTV are also affected by many other factors such as viable tumor cell numbers and intra-tumor heterogeneity [28]. Thus, our research provides a noninvasive method using PET/CT scan to examine tumor cell proliferation.

Chemosensitivity is a vital prognostic factor for ovarian cancer. Previous studies have attempted to earlier estimate chemosensitivity during ovarian cancer neoadjuvant chemotherapy (NACT). Tuulia et al. [29] found that over 57% of SUVmax reduction after four cycles NACT was associated with a better histopathological response. Avril et al. [30] further indicated that PET/CT after the first cycle of NACT could earlier predict chemotherapy response instead of PET/CT after three cycles. Compared to those above, our study showed that we can predict treatment sensitivity before chemotherapy for the first time. According to the SUVmax value, we might select the proper patients for NACT or another individual treatment in a future clinical study.

Previous studies have also shown the promising prognostic significance of PET-derived parameters for ovarian cancer [6, 7, 12, 30]. Yamamoto et al. [12] demonstrated that MTV could, but SUVmax could not, serve as a prognostic factor for PFS in primary ovarian cancer patients. Konishi et al. [6] showed that SUVmax represented a prognostic factor in clear cell adenocarcinoma. Our study failed to find the prognostic value of SUV and MTV. This may be due to the limited number of patients and short follow-up time.

The limitation of our study was that it was a retrospective study with potential recall bias. More patient enrollment and further follow-up are needed in the future to strengthen our findings.

Notwithstanding its limitation, our study involved a homologous group of patients with high-grade serous ovarian cancer, up to 93.8%. Moreover, we focused on ovarian cancer proliferation and chemosensitivity prediction using PET/CT, which is vital but was always neglected by previous studies.

Conclusions

In conclusion, our study suggests that preoperative SUVmax of 18F-FDG PET/CT has a predictive value on proliferation and chemosensitivity after debulking surgery in epithelial ovarian cancer patients noninvasively. Future prospective cooperative multicenter studies are needed to further identify the predictive significance of PET/CT in ovarian cancer.