P1073: A NEW SCORING SYSTEM TO PREDICT SURVIVAL IN ELDERLY ADVANCED STAGE HODGKIN LYMPHOMA PATIENTS: STUDY BY TURKISH SOCIETY OF HAEMATOLOGY LYMPHOMA ACADEMY WORKING GROUP
Makaleler > Tam Makale · 2023
Yazarlar: Özgür Mehtap, Tayfur Toptas, Mehmet Sinan Dal, Güner Hayri Özsan, Nigün Sayinalp, Güray SAYDAM, Mehmet Ali Uçar, Hakki Onur Kirgizlar, Ozan Salim, Atakan Tekinalp, Fahir Özkalemkaş, Funda Pepedil Tanrikulu, Olga Meltem Akay, Emrah Kiliçaslan, Semra Paydas, Sinem Civriz, Mehmet Yilmaz, Volkan Karakuş, Fatma Geçgel, Tahir Darçin, Elçin Erdoğan, Erkin Çinar, Fatma KEKLİK KARADAĞ, Ünal Ataş, Vildan Gürsoy, Salih Sertaç Durusoy, Elif Birtaş Ateşoğlu, Anil Tombak, Nurhilal Büyükkurt, Muhit Özcan, Fevzi Altuntaş, Ahmet Burhan Ferhanoğlu
HemaSphere, Uluslararası, Wos, 2023, cilt 7, ss e566791c-e566791c
DOI: 10.1097/01.hs9.0000971188.56679.1c
Özet
Topic: 17. Hodgkin lymphoma - Clinical Background: Scoring methods are used to predict the prognosis of patients with cHL These predictive scoring methods do not, however, primarily apply to elderly patients. New prognostic scoring (PS) method is required for aged patients due to their clinical traits, treatment outcomes, and prognosis. Aims: We aimed to create a new PS system in elderly cHL patients to predict their survival. Methods: Retrospective data evaluated that collected from inventory of elderly patients with cHL, which comprises a total of 97 variables. We hypothesized that a significant number of patients die within the first year of treatment. We used the derivation cohort (DC) for model building. Dependent variable was overall survival at 12-month. Different variations of the independent variables were tested in the univariate analyses. We built the Cox proportional hazards regression models by a backward stepwise variable inclusion and exclusion procedure. Variables with a p-value <0.10 were included in multivariate analyses. The least significant variables were removed one after one. Cox–Snell residuals were used in assessing overall model fit. Schoenfeld and scaled Schoenfeld residuals were used for checking and testing the proportional-hazards assumption. We transferred the variables, which were the predictors of 12-months death in the stratified Cox model to the clinical scoring system. We used β coefficients of the predictors in the stratified Cox model to assign score weights. The final algorithm was termed as HEDEL12 (Hodgkin’s lymphoma early death in elderly within 12 months). We used the DC as an internal validation cohort (IVC). Both internal and external validations (EVC) were used for model validation. The discriminative ability of the novel scoring system was quantified with Harrel’s concordance (C) statistic. The calibration is the accuracy of prediction of survival or failure probabilities at any time after the time origin. The model calibration was assessed by plotting the predicted against the observed survival. We performed decision curve analysis to determine clinical usefulness and net benefit of the scoring system. Results: 333 patients from 18 centers were included in the analyses. A total of 286 patients fulfilled the inclusion criteria. In DC 38 of 178 patients died within the first 12 months. Median time to death was 5.2 months (95% CI: 4.4 to 6 months). Overall survival (OS) proportion at 12-month was 78.6% (95% CI: 71.8% to 83.9%). Four variables those were independent predictors of HEDEL12: female sex, advanced-stage, low albumin levels (<3.5 g/dL), and ECOG performance scores 2-4. Although the scaled Schoenfeld residuals showed a p-value of 0.15 in global test, the variable “advanced-stage” did not satisfy the proportional hazard. For this reason, we stratified this variable. In the IVC, 12-month failure rates for HEDEL12 scores 0-1, 2, and 3 were 10.2%, 30.0%, and 88.9% respectively. Since there was only nine and seven patients with a HEDEL12 score of 3 in internal and external validation cohorts, respectively, we decided to merge the scores 2 and 3. OS at 12-months were 89.8% and 62.2% for the HEDEL12 scores 0-1 and 2-3, respectively. In the EVC, OS rate at 12-months were 91.0%, 65.2%, for the HEDEL12 scores 0-1 and 2-3, respectively (Figure). HEDEL12 seemed to be perfectly calibrated and yielded an Harrel’s C-statistics of 0.67 and 0.68 for IVC and EVC respectively. Finally, the decision curve analysis revealed that use of the HEDEL12 score had superior net benefit from a threshold probability of 10% or higher for both cohorts. Summary/Conclusion: The HEDEL12 scoring is useful in predicting the survival of older advanced stage cHL patients. Using this scoring, classical chemotherapy can be avoided in high-risk patients.Keywords: Elderly, Hodgkin’s lymphoma