Ovarian cancer (OC) is a gynecological malignant tumor with insidious onset, rapid development, and poor prognosis (1). Worldwide, 313,959 women were diagnosed with OC, and 207,252 new deaths due to this disease occurred in 2020 (2). In China, there were ~25,000 new cases and 22,000 new deaths in 2015 (3). Fewer than half of affected women survive beyond 5 years after diagnosis because of the predominance of aggressive high-grade serous carcinomas and the absence of specific early symptoms and effective early detection strategies (1). Emerging evidence suggests that maximal cytoreductive surgery (4), statin use (5), obesity (6), and vegetable consumption (7) might affect OC survival.

Previous research has suggested that of vegetables, particularly cruciferous vegetables (CVs), might increase OC survival (8). CVs come from plants in the Cruciferae and Brassicaceae families, which account for a large proportion of vegetables, e.g., broccoli, brussels sprouts, cabbage, cauliflower, collard greens, kale, kohlrabi, mustard, rutabaga, turnips, bok choy, and Chinese cabbage (9). Isothiocyanates (ITCs) are formed from the hydrolysis of glucosinolates in plants through the action of the enzyme myrosinase and the gut microbiota, and they are abundant in CVs (10, 11). Different glucosinolates in CVs produce distinct ITCs (11). Some ITCs are volatile and become hydrolyzed at higher cooking temperatures (12, 13). Previous epidemiological studies have indicated that CVs are inversely associated with several health outcomes including cardiovascular disease (14), type 2 diabetes mellitus (15), gastrointestinal cancer (16, 17), breast cancer (18), and reproductive cancers (19, 20). Several epidemiological studies have also shown that ITCs might be associated with various cancers (21–23), diabetes (24, 25), and cardiovascular disease (26, 27). Notably, only three observational studies have focused on the association between pre-diagnosis CV intake and OC mortality; however, these findings were inconsistent. For example, two studies found that a diet high in CVs might help improve OC survival (8, 28). However, one study on 811 patients with OC found no association between CV intake and OC survival (29). More importantly, no study has investigated the relationship between pre-diagnosis ITC intake and OC prognosis.

To provide more evidence indicating the relationships of pre-diagnosis CVs and ITC intake with OC survival, we conducted a hospital-based OC follow-up study in China. To our knowledge, this study is the first large-scale prospective study exploring this topic among Chinese women with different dietary habits, as compared with Europeans and Americans.

The distribution of socio-demographic and clinical characteristics of 703 patients with OC according to tertiles of total CVs is presented in Table 1. The median (interquartile) age at diagnosis was 53.0 (48.0–60.0) years, and no significant difference was observed among the three groups. During a median follow-up of 37.2 months (interquartile: 24.7–50.2 months), we observed 130 deaths, and patients with a higher CV intake had longer follow-up times (p < 0.05). In addition, patients with a higher intake of CVs tended to consume more meat, green leafy vegetables, allium vegetables, fruits, beans and bean products, tea, and total energy. Furthermore, the proportion of postmenopausal women was significantly higher among patients with greater CV intake. Table 2 shows that patients with a non-serous histological subtype, later FIGO stage, and larger residual lesions had poorer survival in this study.

The associations of total CV as well as ITC intake with the overall survival of OC are described in Table 3. The highest tertile of total CV intake was associated with better survival than the lowest tertile intake (HR = 0.57, 95% CI = 0.33–0.98), and a linear trend was also evident (p trend < 0.05). As shown in Figure 2, the OC survival was better in patients with the highest tertile of CV intake than those in the lowest tertile. Moreover, a similar protective effect was observed for ITCs from CV intake (HR _T3VS.T1 = 0.59, 95% CI = 0.36–0.99), although we did not observe a significant linear trend (p trend = 0.06). Analyses of specific species of CVs indicated that the significant protective effect of Chinese cabbage, kohlrabi, and pakchoi was consistent with the main findings (Supplementary Table 1).

The relationship between total CV intake and overall survival in patients with OC was evaluated across potential effect-modifying variables, such as menopausal status, residual lesions, histological subtype, and BMI. The better survival advantages associated with the highest total CV intake were present only in postmenopausal patients, patients with serous histological subtype, or patients with BMI <25 (Figure 3). Additionally, we examined the influence of ITC intake on overall OC survival across the aforementioned potential effect-modifying variables that generated similar patterns (Supplementary Figure 1).

The present findings from this prospective cohort study of 853 Chinese women diagnosed with OC supported that pre-diagnosis CV and ITC intake might enhance the survival of patients with OC. Additionally, we observed significant findings in the analysis of specific species of CVs including Chinese cabbage, kohlrabi, and pakchoi.

Evidence from previous studies evaluating pre-diagnosis CV intake in relation to mortality among OC survivors has been limited and inconsistent. Our findings were consistent with some previous results. For example, Dolecek et al. (28) have reported that higher pre-diagnosis CV intake was significantly associated with longer survival in 341 patients with OC (HR = 0.56, 95% CI = 0.37–0.84). In addition, Nagle et al. (8) have presented findings from 609 women with invasive epithelial OC in Australia, in which pre-diagnosis CV consumption was inversely associated with total mortality (HR = 0.75, 95% CI = 0.57–0.98). In contrast, Playdon et al. (29) have reported a null finding (high to low intake, HR = 0.99, 95% CI = 0.77–1.29) in the results of the Australian OC study (n = 811), which recruited 18–79 year old patients with OC between January 2002 and June 2006. However, interestingly, they observed a trend toward lower mortality risk with higher CV intake when the analysis was restricted to high-grade serous OC cases. These inconsistent findings might be attributable to several reasons, including different demographic and clinical characteristics of patients with OC, exposure assessment, dietary habits, and potential confounders adjustment. For example, Playdon et al. included patients with more advanced OC (III–IV: 71.0%) than our study (III–IV: 48.1%). The diagnostic age composition of patients in our study ( ≤ 50 years: 36.7%) differed from that in the study of Playdon et al. (<50 years: 18.0%). Additionally, although both these studies used an FFQ to collect the diet information, the specific species of CVs differed among them. Of note, the consumption of CVs also differed among these studies. For example, the lower limit of the highest category in the studies of Dolecek et al., Nagle et al., and Playdon et al., was three servings/week (34.28 g/day), 0.83 servings/day (66.40 g/day), and 1.5 servings/day (120.00 g/day), respectively (1 serving equals to 80 g) (39). In contrast, the lower limit of the highest category for CV intake was 70.20 g/day in the present study. Notably, previous evidence has suggested that consumption of CVs differs among Chinese (>100 g/d), Australian (40–80 g/d), and American (25–30 g/d) populations (40). Furthermore, the difference in the potential confounding adjustments among these studies should not be ignored.

The inverse associations of CV and ITC intake with OC mortality are biologically plausible. CVs are good sources of glucosinolates, which can be hydrolyzed by the plant enzyme myrosinase into biologically active compounds such as ITCs. Findings from previous in vivo or in vitro studies have shown that ITCs have cancer-preventive potential (41) through the key mechanisms of inducing carcinogen detoxification and elimination through elevated NRF2 signaling (42, 43), enhancing DNA damage repair (44, 45), and eliminating cancer stem cells (46). Different types of ITCs can also achieve anti-OC ability by inhibiting cell proliferation (47–49) and enhancing cisplatin sensitivity (50, 51). Additionally, CVs contain antioxidants, such as carotenoids, vitamin C, and vitamin E. Antioxidants can counteract the effects of reactive oxygen species, including initiation lipid peroxidation, cell damage, disruption of cell signaling, and increased viral replication and expression, all of which are pivotal events in carcinogenesis (52, 53).

This study has several crucial strengths. First, this is the first investigation to report a survival advantage among patients with OC with higher intake of pre-diagnosis CVs and related ITCs in China. Second, the OOPS study was a well-designed prospective cohort study with high baseline survey participation rates and follow-up retention rates, thus minimizing the potential for recall bias or selection bias. Further, we used a validated FFQ to evaluate dietary information through face-to-face interviews, thus providing stable estimates of CV intake. Third, a wide range of potential confounders associated with OC prognosis were strictly adjusted for, such as residual lesions, histological type, and FIGO stage. Fourth, Asian populations habitually consume large amounts of CVs and other plant-based foods, thus enabling hypotheses associated with the underlying health properties of these foods to be investigated specifically (54).

However, potential study limitations also must be considered. First, FFQ was used to evaluate CV intake of OC patients. The standard size of food share was applied to calculate food intake, but the intake of patients could not be accurately assessed. Thus, information bias is inevitable. However, we used a highly reproducible validated FFQ and selected highly trained and skilled researchers to collect dietary information. Additionally, spearman correlation coefficients and intraclass correlation coefficients of reproducibility for CV were 0.61 and 0.53, respectively. Second, compared with the average intake of CVs (~22.6 g/d) in Western populations (55), the intake in Chinese populations exceeds 100 g/d (40), thus potentially limiting the generalizability of our findings to the consumption of CVs in Western populations. Furthermore, ITC exposure levels are influenced by various cooking methods (9). Boiling CVs leads to a 30–60% loss of complete glutathione S-transferases, owing to leaching and thermal degradation (56). Nevertheless, none of the studies have distinguished CV intake according to food preparation method. Hence, further studies must focus on whether the inverse correlation of CV intake might be influenced by the cooking method. Third, the relationship between CV intake and progression-free survival in patients with OC was not evaluated in our analysis; however, owing to the poor prognosis and short post-progression survival period of this disease, the progression-free survival was similar to overall survival (57, 58). In addition, according to previous similar research (8, 29, 59), we analyze all-cause deaths rather than OC-specific deaths in the present study, because research shows that there is little difference between them (60). Last, although we have comprehensively adjusted for potential confounding covariates in multivariate models, these findings still may be affected by unknown or unmeasured confounders including different clinical treatment and types of mutation (61, 62). Additionally, residual confounding from unmeasured confounders as well as the assessment of dietary intake may also influence the results, since these variables were assessed base on self-administered questionnaires.

CVs are commonly consumed worldwide, and they are rich in nutrients, such as ITCs and antioxidants. In the present study, increasing pre-diagnosis consumption of CVs was strongly associated with better survival in patients with OC. This protective effect was similar in the analysis of ITCs from CVs. Nevertheless, further studies with longer follow-up periods are needed to confirm our findings.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

The studies involving human participants were reviewed and approved by the Institutional Review Board of the Ethics Committee of Shengjing Hospital of China Medical University, Shenyang, China. The patients/participants provided their written informed consent to participate in this study.

Q-JW, Y-HZ, and T-TG: contributed to the study design. SG, X-QL, SY, and T-TG: collection of data. Y-FW and F-HL: analysis of data. Y-FW, Y-YH, F-HL, Z-YW, H-YW, SZ, ML, and Q-JW: wrote the first draft of the manuscript and edited the manuscript. All authors have read and approved the final manuscript.

This work was supported by the National Key R&D Program of China (No. 2017YFC0907404 to Y-HZ), the Natural Science Foundation of China (No. 82073647, 81602918 to Q-JW, and 82103914 to T-TG), China Postdoctoral Science Foundation Funded Project (No. 2018M641752 to Q-JW), LiaoNing Revitalization Talents Program (No. XLYC1907102 to Q-JW), Shenyang high level innovative talents support program (No. RC190484 to Q-JW), and 345 Talent Project of Shengjing Hospital of China Medical University (Q-JW).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.