Enlarged cardiophrenic lymph nodes predict disease involvement of the upper abdomen and the outcome of primary surgical debulking in advanced ovarian cancer

Abstract Introduction The outcome of ovarian cancer patients is highly dependent on the success of primary debulking surgery in terms of postoperative residual disease. This study critically evaluates the clinical impact of preoperative radiologic assessment of the cardiophrenic lymph node (CPLN) status in advanced ovarian cancer. Material and methods Baseline CT scans of 178 stage III and IV ovarian cancer patients were retrospectively reviewed by two independent radiologists. CPLN enlargement defined at a short‐axis ≥5 mm was evaluated for its prognostic value and predictive power of upper abdominal tumor involvement and the chance of complete intra‐abdominal tumor resection at primary debulking surgery. Only patients without surgically removed CPLN were eligible for this study. Results Enlarged CPLNs were detected in 50% of patients and correlated with radiologically suspicious (P = .028) and histologically confirmed (P = .001) paraaortic lymph node metastases. CPLNs ≥ 5 mm were associated with high CA‐125 levels at baseline and revealed independent prognostic relevance for progression‐free survival (hazard ratio [HR] 2.14, 95% confidence interval [CI] 1.33‐3.42) and overall survival (HR 2.18, 95% CI 1.16‐4.08). Noteworthy, patients with enlarged CPLNs nonetheless benefit from complete intra‐abdominal tumor debulking in terms of an improvement in progression‐free survival (HR 0.60, 95% CI 0.38‐0.94) and overall survival (HR 0.59, 95% CI 0.35‐0.82). Enlarged CPLNs correctly predicted carcinomatosis of the upper abdomen in 94.6%. A predictive score of complete tumor debulking, termed CD‐score, which integrates, beside a CPLN short axis <5 mm, an ascites volume <500 mL, and CA‐125 levels <500 U/mL at baseline, correctly predicted complete intra‐abdominal debulking in 100% of patients. Conclusions CPLNs ≥5 mm predict upper abdominal tumor involvement. The application of the CD‐score predicted complete macroscopic tumor resection at primary surgery in all of the patients. Although, CPLN pathology suggests extra‐abdominal disease, we consistently demonstrated that patients nonetheless benefit from complete intra‐abdominal tumor resection.


| INTRODUC TI ON
Ovarian cancer is the most lethal gynecologic malignancy in western countries. 1 Mainly diagnosed at an advanced stage, prognosis is fundamentally influenced by residual disease after debulking surgery. 2 The achievement of optimal cytoreduction is highly influenced by the extent of carcinomatosis, whereby tumor spread in the upper abdomen is a major obstacle to complete macroscopic tumor resection. 3 In this context, high-resolution imaging modalities are used in the preoperative evaluation of tumor spread in order to predict complete tumor resectability. 1 Over the last decades, extensive pelvic and paraaortic lymphadenectomy up to the left renal vein has been performed to remove potentially cancer-affected retroperitoneal lymph nodes. However, data from the LION trial did not show any benefit of systematic lymph node dissection in advanced stage ovarian cancer in the case of radiologic unsuspicious nodes. 4 Despite the intention to remove occult cancer by performing systematic lymphadenectomies, virtually no attention has been paid to the frequently observed enlarged cardiophrenic lymph nodes (CPLNs). 5,6 Anatomically, the cardiophrenic region is a fat-filled space between the mediastinum, heart base, diaphragm and chest wall. Small visible lymph nodes in this area can be physiological but there usually are fewer than two, each with a diameter of <5 mm. 7 The most classical route of dissemination of epithelial ovarian cancer is the spread of exfoliated free-floating cancer cells throughout the abdominal cavity via the physiological peritoneal fluid, leading to peritoneal carcinomatosis. The major portion of the lymph drainage to the CPLNs occur via sub-peritoneal plexuses located adjacent to the diaphragm. Holloway et al found a correlation between peritoneal metastases and the enlargement of paracardiac lymph nodes. 6 In RECIST 1.1, lymph nodes, independently of their location, are considered to be of pathologic relevance when their short-axis is >10 mm. 8 Nevertheless, several groups have demonstrated an adverse prognostic impact on survival of ovarian cancer patients exhibiting CPLNs with >5 mm short-axis. 6,9,10 Thus, the European Society of Urogenital Radiology (ESUR) has defined pathologic enlarged CPLNs at a cutoff a ≥5 mm short-axis dimension. 11 The study presented here aims to evaluate critically the prognostic impact of enlarged CPLNs in advanced stage ovarian neoplasms, especially considering the debulking outcome at primary surgery. In addition, we studied the impact of radiologically assessed pathologic CPLNs to predict carcinomatosis of the upper abdomen and complete macroscopic tumor resection at primary debulking surgery.

| Statistical analyses
According to the European Society of Urogenital Radiology (ESUR) guidelines 11 and a recently published study, 10 we defined pathologically enlarged CPLNs at a cutoff ≥5 mm short-axis dimension. All statistical analyses were performed using the SPSS® Statistics software version 24 (IBM, Armonk, NY, USA). Differences in survival were assessed using the Kaplan-Meier method with log-rank test and Cox's proportional hazard models for uni-and multivariate analyses. Association of CPLNs ≥5 mm and clinicopathological parameters was evaluated using Mann-Whitney U or Chi-square tests. Statistical significance was defined as P < .05.

| Ethical approval
The utilization of patient data for study purposes was handled ac-

| Radiological detection of cardiophrenic lymph nodes and clinicopathological parameters
In a cohort of 178 advanced ovarian cancer patients, preoperative CT scans revealed CPLNs of any size in 90.4% (n = 161) of patients; 50.0% (n = 89) of patients exhibited lymph nodes ≥5 mm in short-axis.
In 48.9%, the dominant lymph node was located at the right side of

| Prognostic impact of enlarged cardiophrenic lymph nodes
CPLNs ≥5 mm were associated with impaired progression-free and overall survival ( Figure 1A,B). In addition, the CPLN status proved to be of independent prognostic relevance in Cox's regression models including the age at diagnosis, residual disease after debulking surgery, CA125 level, FIGO stage and retroperitoneal lymph node positivity as covariates ( Table 2) (Figure 2A,B).

| Predictive value of enlarged cardiophrenic lymph nodes for upper abdominal disease
Intraperitoneal carcinomatosis was radiologically evident in 84.8%   (Table 3).

| D ISCUSS I ON
In the cohort examined here, we found enlarged CPLNs (short-axis ≥5 mm) in 50% of patients. In the literature, detection rates of between 11% and 62% are described depending on the discriminatory diameter used and the patient cohort investigated. 6,[9][10][11][14][15][16][17] A stepwise increase of radiologically used CPLN short-axis from ≥3 to ≥10 mm yields to a substantial decrease in the detection rate of potentially pathological CPLNs. 14 This effect is corroborated by several studies and was confirmed by our own data (Table S1). The histological confirmation rate of enlarged CPLNs ranges between 85% and 95% 10,[17][18][19]22 and is obviously independent of different radiological short-axis diameters (Table S1) was demonstrated in 85%-95%. 10,[17][18][19]22 Nonetheless, the possibility of a reactive lymph node enlargement based on "work hypertrophy" in the case of ascites has to be kept in mind. 9 Regarding the latter concern, our data did not reveal a positive association between enlarged CPLNs and the amount of ascites, or the occurrence of pleural effusions.

| CON CLUS ION
The data presented here of an inverse association of enlarged CPLN with complete resection rates and patient survival confirm the recently published data by Prader et al. 10 In addition, we demonstrate here that a preoperative radiological evaluation of the CPLN status anticipates tumor involvement of the upper abdomen, which is especially useful in those patients with miliary carcinomatosis missed by conventional imaging. Beyond this, we provide for the first time an easy-to-perform clinical score to predict the chance of complete surgical debulking, which could help clinicians in tailoring adequate therapy strategies in ovarian cancer. In addition, we demonstrate that patients with enlarged CPLNs nonetheless benefit significantly from all possible surgical efforts translating into a complete intraabdominal tumor resection.

CO N FLI C T O F I NTE R E S T
The