A large number of ovarian cancer patients diagnosed with advanced FIGO stage (III-IV). Generally speaking, patients with stage IVB ovarian cancer have an unfavorable prognosis, with a median OS of 25.2-30.0 months [11]. However, several studies had shown that the survival of ovarian cancer with distant lymph node metastasis as the only evidence for stage IVB was significantly better with a median OS of 39.0-41.1 months [11, 14]. Some authors believed the existing staging system couldn’t fully and accurately predict the biological behavior and prognosis of stage IV ovarian cancer [15]. Patients with parenchymal organ metastasis usually had poor prognosis and should be separated from those who had relatively better prognosis [16]. As for the reason why IVB patients with distant lymph node metastasis had better prognosis, some scholars from MD Anderson Cancer Center [10] believed that this might because of their lower tumor burden in abdominal and pelvic cavity, which lead to higher possibility to achieve optimal abdominopelvic debulking [4]. As evidence, study revealed that for IVB patients with distant lymph node metastasis, the median survival of patients with no macroscopic lesions in omentum was as long as 120 months, while it was only 24 months in patients who had massive lesions in omentum [10]. Some studies also believed that this might because of those patients’ better general condition and tolerance to aggressive debulking and multi-cycle chemotherapy [17]. Our study confirmed the heterogeneity in supraclavicular metastatic ovarian cancer with a large sample size. In addition, when we focused on the group with better prognosis, we found optimal debulking of abdominopelvic cavity in primary surgery played a significant role.

The prognosis of patients with “continuous-metastasis type” was much better

In our study, patients with supraclavicular lymph nodes metastasis accompanied by metastatic PALNs in imaging (PET-CT/CT/MR) before PDS or NACT were defined as “continuous-metastasis type” while patients without metastasis in PALNs defined as “skip-metastasis type”. We found the median survival of the patients of “continuous-metastasis type” was almost twice as the data of patients with “skip-metastasis type”. Patients with “continuous metastasis-type” accounted for 86.3% of the whole cohort in our study, which was close to the proportion (87.0%) of PALNs metastasis in stage IV ovarian cancer reported in previous literatures [18]. Since some patients did not receive preoperative PET-CT, but only received abdominopelvic CT/MR, chest X-ray and supraclavicular lymph node biopsy, they could not be precisely evaluated for preoperative intrathoracic lymph node metastasis. In that case, we did not adopt intrathoracic lymph nodes as a criterion for judging continuous or skip metastasis. Generally, ovarian cancer that metastasizes to supraclavicular lymph nodes mainly reaches the retroperitoneal para-aortic lymph nodes through the suspensory ligament or pelvic lymph nodes, and then spreads to the upper part of the diaphragm then to the supraclavicular region, through the diaphragmatic and retroperitoneal lymphatic drainage. This "continuous-metastasis" pattern was known as the classic lymphatic drainage pattern of ovarian cancer. However, in some rare cases, cancer cells could metastasize to isolated lymph nodes in distance through blood, showing a pattern of skip metastasis [7]. In additions, peritoneal tumors can drain through diaphragmatic lymphatic vessels to major veins above the diaphragm, which also leads to skip metastasis [1]. Both of the two ways usually imply high burden of tumors and unfavorable biological character of the tumor. Therefore, we speculated that these two anatomical theories might explain why patients defined as “skip-metastasis type” was associated with worse prognosis.

Among the 44 patients with continuous metastasis, suspected PALNs were found in 32 patients during surgery, 22 of whom underwent paraaortic lymph node dissection (PAND) and 16 had no residual metastatic PALNs. Nineteen patients were confirmed with metastatic PALNs by histology. As for the other 10 cases with suspected PALNs, the reason why they didn’t undergo PAND was that their metastatic PALNs were too fixed to be dissected, which means PALNS were closely adhered to main vessels or fixed to the retroperitoneum. Suspected PALNs were not found in 12 continuous metastasis patients during IDS, which might be attributed to the NACT. Therefore, none of them underwent PAND during IDS. Based on the results of this study, we strongly recommend optimal debulking including lymphadenectomy in PDS, while the immunity of lymphadenectomy for those without suspected PALNs in IDS requires further studies.

Optimal abdominopelvic debulking had prognostic benefit for “continuous metastasis type” patients

The independent factors that reported to affect the prognosis of stage IV ovarian cancer varied, which included: age, physical status, location of metastatic lesions, the volume of ascites, residual tumor after surgery, radical surgery, chemotherapy regimen, etc [3,4,5,6,7,8,9, 19,20,21,22,23,24]. However, the most widely recognized prognostic factor was achieving optimal debulking in primary surgery [4,5,6,7,8,9, 19,20,21,22,23]. For stage IV patients, to achieve optimal debulking of no residual tumor in entire body sometimes means extremely aggressive and multi-incision surgeries, which may include intrathoracic surgeries [25,26,27,28] and lymph node dissections of neck [13, 29]. However, the increased risk of complications and the reduced quality of life should not be ignored in such extensive cytoreductions [26]. At the same time there are different opinions. Some studies proved even with distant metastasis, the main prognostic factor for stage IV ovarian cancer patients was abdominopelvic tumor [24]. Besides, progress in adjuvant therapy could also provide novel treatments for distant metastases postoperatively. In such scenario, optimal debulking of the pelvic and abdominal cavity still benefits patients with distant lymph node metastasis. In our whole cohort, we found that the prognosis of patients who had optimal abdominopelvic debulking was better than those who had suboptimal abdominopelvic debulking, but the difference was not statistically significant (HR = 2.47, 95% CI: 0.86-7.13). We figured the prognostic effect of distant lymph node metastasis might mask the potential benefit of optimal debulking. Therefore, we analyzed the impact of optimal abdominopelvic debulking on prognosis in the subgroup of patients with “continuous-metastasis type”. Our data showed that in patients with “continuous-metastasis type”, the OS of those who achieved optimal abdominopelvic debulking was 55.3 month, significantly longer that the OS of those who were not optimally abdominopelvic debulked (42.3 months, p= 0.034). There were studies in which optimal systemic debulking of entire body was done for IVB patients, and the survival data they reported were 25-55 months. [5,6,7,8,9, 19]. Although head-to-head comparisons were not possible, the survival of optimally abdominopelvic debulked patients in our study were not inferior to the survival data reported in literatures. Several theories might explain the prognostic benefit of optimal abdominopelvic debulking. Firstly, most ovarian cancer recurrences were located in abdominopelvic cavity, rather than distant lymph node region [12, 30]. Secondly, compared with the supraclavicular region, recurrences or residual tumors in abdominopelvic cavity were more likely to result in fatal complications, such as intestinal obstruction, cachexia and infection, while supraclavicular tumors seldom cause serious symptoms [3, 24]. In addition, adjuvant therapy might also effectively control the supraclavicular lesions [12]. Optimal systemic debulking of the entire body for stage IV ovarian cancer was never easy, since the rates reported in previous studies never reached 50% [5,6,7,8,9, 19]. Therefore, for stage IV ovarian cancer with supraclavicular metastasis, especially those with PALNs metastasis (“continuous-metastasis type”), we suggested the goal of primary cytoreduction might be no residual tumors in abdominopelvic cavity.

NACT+IDS vs. PDS

Our study also found that prognosis of patients received PDS was not inferior than those received NACT + IDS (HR = 1.97, 95% CI:0.68-5.70). In fact, the debate on the timing of surgery for stage IV ovarian cancer patients never stopped. In previous retrospective studies comparing the prognosis of IDS and PDS in IVB ovarian cancer patients, results in which PDS were better than, worse than, or equal to IDS were reported separately [4, 12, 31]. Two recent large prospective clinical trials [32, 33] showed that there was no significant difference in survival between stage IV patients received NACT + IDS or PDS, which was similar to our results. Though patients received PDS may had higher rates of complication, our data revealed that if the complications were managed by experienced gynecologic oncologists, patients could achieve equal or even better prognosis than IDS. We believed the main possible advantage of PDS could be, it helped to clarify the extent of tumor, avoid the omission of small lesions after chemotherapy and reduce the resistance to chemotherapy. Larger prospective studies are encouraged to further clarify this issue.

BRCA mutations

Another factor that had been reported to significantly improve the prognosis of stage IV ovarian cancer was BRCA mutations [30]. Since the genetic testing could not be covered by medical insurance and was only advocated in clinical practice in the recent years in China, 27 patients (52.9%) in our cohort didn’t have records about BRCA mutations, while only 7 patients (13.7%) were recorded as noncarriers. Considering the non-response bias in BRCA status (some “unknown” were actually “negative”), the actual rate of g/sBRCAm in our study was at least 33.3% (17/51), which was already higher than the reported BRCA rate of 28.5% in the largest study of Chinese ovarian cancer patients [34] and 5%-29% mostly in patients from white background [34,35,36,37,38,39,40]. Whether it is because patients with BRCA mutations are more prone to supraclavicular lymph node metastasis, or because there is a higher BRCA mutation rate in patients with supraclavicular lymph node metastasis, it is worthy of further studies. It is worth mentioning that we did find 1 patient with gBRCA2 mutation achieved very good OS of 63.4 months. This patient had residual metastatic pelvic lymph nodes of 0.5 cm after primary debulking and received PARPi after her last recurrence which lasted for nearly 2 years. Whether the BRCA mutation rate was higher in IVB stage patients with supraclavicular lymph node metastasis, or the treatment benefit of PARPi can offset the defect of suboptimal debulking of leaving lymph nodes remains to be further studied.

Limitations

Although we presented the study on supraclavicular metastatic ovarian cancer patients with the largest sample size, there’s no denying our study had some limitations. As a retrospective study, it was difficult to tell whether the survival advantage was due to the successful surgery, or it was the favorable biological nature of the tumor made the optimal debulking possible. In addition, the distribution of BRCA status in this group of patients should be studied in further follow-up. Another limitation was the small sample size, especially the number of patients in the “skip-metastasis type” group due to the rare condition. Since the number of ovarian cancer patients with supraclavicular lymph node metastasis is relatively small, prospective multicenter studies should be encouraged to further verify the findings of this study.