Objective. The objective of this study was to verify the correlation between p53 immunostaining at initial diagnosis and at positive reassessment after completing platinum-based chemotherapy and to assess prognostic differences between patients whose tumors display positive immunostaining versus those that have negative immunostaining at such reassessment. Methods. This study made use of samples from patients entered into a prospective randomized study of the Southwest Oncology Group (SWOG 8835) that treated patients with minimal residual disease at second-look laparotomy with either intraperitoneal (ip) mitoxantrone or fluorodeoxyuridine (FUDR). Unstained slides from tumor obtained at the initial diagnosis and at reassessment were retrospectively requested from individual institutions. The degree of nuclear staining was determined using the anti-p53 mouse monoclonal antibody Pab1801 and previously published techniques, with a cutoff of 10% or more staining of tumor cell nuclei for a positive result. Cox model regression analysis was performed for overall survival and progression-free survival, with p53 status, ip treatment, and baseline CA125 as independent variables. Results. p53 determination was feasible in 22 patients both at diagnosis and at the second-look samples; 9 additional patients had only either sample available. Since concordance between the 10 negative and 12 positive immunostained samples was 100%, all 31 patients were considered in the Cox model. The death hazard ratio of p53-positive versus p53-negative patients was 4.18 (two-sided P value of 0.006). Conclusion. p53 immunostaining at second-look laparotomy correlates with the immunostaining at diagnosis. In this series confined to patients with minimal residual disease after initial therapy subjected to second-line intraperitoneal treatment, it appears to identify a poor prognostic (positive) subset for survival.