Validation of Childhood Hodgkin International Prognostic Score (CHIPS) for Predicting Event-Free Survival in Intermediate and High-Risk Hodgkin Lymphoma

Abstract

Background: Improving risk stratification of patients with Hodgkin lymphoma (HL) allows for optimization of treatment allocation and minimization of late effects when possible. The Childhood Hodgkin International Prognostic Score (CHIPS) was developed as a predictive model for event-free survival (EFS) in pediatric and adolescent HL using clinical data at diagnosis from patients with intermediate-risk HL treated on Children's Oncology Group protocol AHOD0031 (NCT00025259) with doxorubicin, bleomycin, vincristine, etoposide, prednisone, cyclophosphamide (ABVE-PC) chemotherapy and response-adapted radiation. Stage IV, large mediastinal mass, albumin <3.5 g/deciliter, and fever were identified as independent predictors of EFS and assigned one point each, resulting in CHIPS ranging from 0 to 4. CHIPS was highly predictive of EFS in this patient cohort, but has not been validated in high-risk patients or in patients treated with other therapeutic regimens. We aim to validate the CHIPS as a risk stratification tool for patients with intermediate or high-risk HL treated with Stanford V chemotherapy by the Pediatric Hodgkin Consortium (PHC). Methods: The PHC trial HOD99 (NCT00145600) enrolled patients on the high-risk arm with stage IIB, IIIB, or IV disease <22 years (n=123). HOD05 (NCT00352027) enrolled patients with intermediate-risk HL with stage IB, IA or IIA with “E” lesions, ≥3 nodal sites, or bulky mediastinal adenopathy, or IIIA disease <22 years (n=49). All patients received 12 weeks of Stanford V chemotherapy with doxorubicin, vinblastine, mechlorethamine, vincristine, bleomycin, etoposide, and prednisone followed by low-dose, response-adapted radiation therapy. All patients with absent B symptoms were deemed afebrile; patients missing documentation of any elements of CHIPS were excluded from the analysis. Results: Our final analysis included a total of 164 patients, 116 with high-risk HL from HOD99 and 48 with intermediate-risk HL from HOD05. One patient was excluded for missing fever data and 7 were excluded due to missing albumin values. Patients were classified into 2 groups based on their CHIPS value: CHIPS 0-1 (n=88, 53.7%) and CHIPS 2-4 (n=76, 46.3%). The EFS for patients in the CHIPS 0-1 and CHIPS 2-4 groups was analyzed using Kaplan-Meier curves (Figure 1). The 2-year EFS estimates for patients with CHIPS 0-1 and CHIPS 2-4 were 95.5% (95% confidence interval (CI) 91.2%-99.9%) and 80.3% (95% CI 71.8%-89.7%), respectively, with a log-rank p-value of 0.003. Similarly, the 4-year EFS estimates for patients with CHIPS 0-1 and CHIPS 2-4 were 93.2% (95% CI 88.0%-98.6%) and 77.6% (95% CI 68.7%-87.6%), respectively, with a log-rank p-value of 0.004. The reduced EFS in patients with higher CHIPS is consistent across different stages and early response assessment (Table 1). Conclusions: CHIPS is highly predictive of EFS in pediatric and adolescent patients with intermediate and high-risk HL treated with Stanford V chemotherapy, as it identifies a subset (CHIPS 2-4) with significantly lower EFS compared to the other subset (CHIPS 0-1). Assessment of CHIPS alongside novel approaches such as circulating tumor DNA and total metabolic tumor volume should be analyzed across trials to allow for further enhancements to risk stratification for pediatric, adolescent and young adult HL.