Background: Poly(ADP-ribose) polymerase inhibitor (PARPi) treatment in ovarian cancer patients after first-line chemotherapy and following the response to platinum-based chemotherapy at relapse is associated with survival benefits. Maintenance therapies can be administered over extended periods, making treatment tolerability assessment essential in optimizing patient outcomes. This cohort study aimed to evaluate the agreement between physician and patient reporting of PARP inhibitor-related toxicities and the rate of underestimation of each symptom considered. Methods: Patients treated with PARPis in the first-line or recurrent setting were included. A specific Patient-Reported Outcomes Common Terminology Criteria for Adverse Events (PRO-CTCAE) questionnaire was generated and administered to the cohort. For each toxicity, agreement between patients and physicians was assessed using Cohen’s kappa and Gwet’s AC1; in addition, the rate of toxicity under-reporting by physicians was calculated. Results: Seventy-seven ovarian cancer patients were included; 39 (50.6%) received PARPis in the first-line setting, while 38 (49.4%) were treated for recurrence. Cohen’s kappa values for agreement between patients and physicians across 12 reported toxicities ranged from 0 to 0.15, indicating poor agreement (κ < 0.20) for all assessed toxicities, with the lowest levels of agreement for decreased appetite (κ = 0), rash (κ = 0.02), headache (κ = 0.00), arthralgia (κ = 0.03), insomnia (κ = 0.03), and fatigue (κ = 0.04). When agreement was assessed using Gwet’s AC1, agreement remained poor to moderate for the majority of the symptoms evaluated. Physician under-reporting rates were higher for nausea (51.9%), rash (57.1%), headache (49.3%), arthralgia (70.2%), insomnia (48.1%), and fatigue (67.5%). Conclusions: Our results underscore the importance of systematically integrating patient-reported outcomes into clinical practice, including in maintenance settings, to ensure an accurate assessment of treatment-related toxicities.