Examining the Selection Criteria of Neoadjuvant Chemotherapy Patients

To identify predictors of neoadjuvant chemotherapy (NAC) and to examine toxicities, dose reduction, interruptions, and second-line chemotherapy MATERIALS AND METHODS: A retrospective chart review of 391 patients with late-stage ovarian cancer diagnosed between January 1, 2004 and December 31, 2010 was conducted. Logistic regression was used to predict chemotherapy type. Cumulative incidence of toxicities, dose reduction, and treatment interruption were calculated using the Kaplan-Meier method. Overall survival was analyzed using time-varying Cox regression models. A competing risk model was used to predict second-line chemotherapy with death as a competing risk.

Older patients were less likely to receive primary debulking (OR 0.710; 95% CI 0.55-0.92, P = 0.0108), as were patients with longer diagnostic intervals. Clear-cell, endometrioid, and mucinous carcinoma were more likely to receive adjuvant treatment than unclassified epithelial (OR 6.964; 95% CI 2.02-24.03, P = 0.0021). Adjuvant patients experienced higher incidence of chemotherapy toxicities (P <0.0001) and treatment interruption (P = 0.016) at 3 months. There was no statistically significant difference in the incidence of chemotherapy dose reduction of >20% in the NAC and adjuvant populations (P = 0.142). Neoadjuvant patients were more likely to require more than one line of chemotherapy ([Subhazard Ratio] = 4.334; 95% CI 2.51-7.50, P <0.0001).

Our study found that patients with shorter diagnostic intervals, more advanced age, and unclassified epithelial histotype were more likely to receive NAC. NAC patients did not experience a higher incidence of chemotherapy toxicities, treatment interruption, or dose reduction. There is treatment selection bias for sicker patients being treated with NAC.