Identification of patient groups at highest risk from traditional approach to ovarian cancer treatment

Abstract

Objective

Define subgroups of patients at highest risk for major morbidity and mortality after a traditional approach of maximal surgical efforts followed by chemotherapy for advanced ovarian cancer (AOC).

Methods

Preoperative health, intra-operative findings and outcomes were assessed in consecutive patients with primary AOC from 4 centers. Initial tumor dissemination was stratified into 3 groups based on volume of disease. Surgery was categorized using a previously described surgical complexity score (SCS). Statistical analysis was directed toward validating a multivariable risk-adjusted model.

Results

576 patients with stage IIIC (N = 447, 77.6%) or IV AOC (N = 129, 22.4%) were analyzed. Age (HR (per year): 1.02; 95%CI: 1.01–1.03), high tumor dissemination (HTD) (HR: 1.73; 95%CI: 1.19–2.56), residual disease (RD) > 1 cm (HR: 2.46; 95%CI: 1.74–3.53), and stage IV (HR: 1.93; 95% CI: 1.51–2.45), independently correlated with OS. We identified a small subgroup of patients who comprised a high-risk group (N = 38, 6.6%) characterized by all of the following characteristics: high initial tumor dissemination (HTD) or stage IV plus poor performance or nutritional status plus age ≥ 75. In this group, high SCS to achieve low RD was associated with morbidity of 63.6% and limited survival benefit.

Conclusions

Optimal management of AOC requires accurate, risk-adjusted predictors of outcomes allowing a tailored approach starting with primary therapy. Complex surgical procedures to render low RD improve survival, and in the majority of cases, the benefits of such surgery appear to outweigh the morbidity. However careful analysis identifies a subgroup of patients in whom an alternative approach may be the better strategy.

Introduction

Standard recommendations for patients with advanced ovarian cancer (AOC) include surgical cytoreduction followed by platinum-based adjuvant chemotherapy [1]. The rationale for surgical cytoreduction finds its first evidence in 1975 with a publication by Griffith [2]. Subsequently, several publications confirmed the role for primary cytoreductive surgery in management of ovarian cancer [3], [4], [5], [6], [7], [8], [9]. Countering this thinking is the concept of the importance of intrinsic tumor behavior which cannot be completely overcome with surgical resection [10], [11], [12], [13]. While undoubtedly initial stage remains an important independent variable, cytoreduction appears to improve survival regardless independent of stage [13], [14]. It is widely recognized that patients with stage IIIC disease and carcinomatosis, or widely disseminated peritoneal disease, carry a worse prognosis than patients without it [6]. But we and others have also shown that surgical cytoreduction can improve survival to equal those with less initial disease volume [5], [6], [14]. However, some patients are poor candidates for aggressive surgical procedures and might be better served by an alternative approach.

An important challenge if we are to maximize overall outcomes for our patients is to better identify the subgroups of patients who are unlikely to benefit from an aggressive surgical approach, thus avoiding unnecessary morbidity and short term mortality which dramatically impact and shorten remaining life of these women and dramatically raise cost of care. This is particularly cogent as many oncologists are employing a neoadjuvant approach which has not been directly compared to aggressive primary surgery. Careful analysis of outcomes can help in the selection of most appropriate candidates for non-standard therapy. We previously demonstrated that age, low pre-operative levels of albumin, poor performance standard were all predictors of postoperative morbidity [15]. Others have recently shown the predictive value of complexity of the surgery and impaired performance status along with obesity and elevated SGOT to be predictors of perioperative morbidity [16]. Thus, the primary question we tried to answer was whether we could identify a subgroup of patients with particularly high risk, unlikely to tolerate complex surgery in a multi-center model to reduce institutional bias. Ultimately, we were able to identify three predictors (age, poor performance or nutritional status, and disseminated disease) that, combined together, constituted a small, but very “high risk group” of patients in which aggressive surgical efforts do not translate into improved survival and are associated with significant negative impact.