Introduction:

Cyclophosphamide is a chemotherapeutic agent with antineoplastic and immunosuppressive properties used for a broad range of indications. Short-term adverse effects include nausea and vomiting, myelosuppression, and hemorrhagic cystitis. Longer-term toxic effects included gonadal effects, and notably, bladder malignancy.

Cyclophosphamide exposure has a dose-dependent relationship with bladder cancer. The characteristics of cyclophosphamide-related bladder cancer have not been well studied outside of case reports and series though. Additionally, surveillance regimens for patients following cyclophosphamide administration are not informed by high levels of evidence or standardized guidelines. Thus, we conducted a systematic literature review of cyclophosphamide-related bladder cancer to better inform survivorship care for patients.

Methods:

Medline, Web of Science, and Cochrane Library were searched using the terms “cyclophosphamide bladder cancer” from inception through August 2020, and studies with individual-patient data were included. Reference lists within each study were examined to ensure inclusion of pertinent studies, and where able to be translated, non-English reports were included. Patient populations in each study were scrutinized for duplicates, and where identified, the most recent study was used.

Data extracted from each paper included: patient gender; indication for cyclophosphamide; presenting symptoms before diagnosis; bladder cancer type; TNM staging; tumor grade; age of onset; latency time (time between initial cyclophosphamide administration and cancer diagnosis); total dosage; treatments received subsequent to diagnosis; vital status and whether death was cancer-related. Due to report heterogeneity, not all data listed above were obtained for each patient. Descriptive statistics and data visualization were performed with Excel.

Results:

We identified 298 patients from 122 studies, with a male predominance (55.1%). The most common malignant indication was lymphoma (23.5%); the most common non-malignant indication were ANCA-associated vasculitides (24.6%). Hematuria and dysuria were most prevalent prior to diagnosis. Conventional urothelial carcinomas (UCs) were most common, followed by leiomyosarcomas and UC with variant morphology. About half of cancers were invasive, and most were high grade. Table 1 summarizes tumor characteristics.

Median age at diagnosis was 56.0 years (IQR 25.0), and median latency time was 10.0 years (IQR 6.9); overall distribution of latency time is summarized in Fig. 1. Median cumulative dose was 108.0 grams (IQR 133.0). Of 132 patients with treatment data, 25 (18.9%) received radiation therapy, 25 (18.9%) received chemotherapy or immunotherapy, and 74 (56.1%) underwent partial or radical cystectomy. Of 126 patients with known vital status, 55 died (43.7%), with 43 of these deaths attributable to cancer.

Conclusion:

We describe the largest pooled analysis of patients with cyclophosphamide-related bladder cancer. These cancers have a propensity for more aggressive behavior and non-urothelial histology. A substantial number of patients with extended latency time of ≥20 years (10.4%) was noted, supporting long-term monitoring of these patients for bladder cancer. Hematuria was the most common symptom prior to diagnosis, making annual urinalysis a reasonable initial screening test, with further workup with cystoscopy and upper tract imaging if hematuria is present.

Few patients in this cohort received systemic therapy and future work should examine its utility in this population and patient or disease factors that may preclude its use in this cohort. Our analysis is limited by an inability to account for confounding factors that may predispose patients to bladder cancer external to cyclophosphamide exposure, as well as incomplete information regarding treatment regimens beyond cumulative dose.

Funding: N/A