ITALIAN VALIDATION OF THE INTERNATIONAL PROGNOSTIC INDEX FOR CHRONIC LYMPHOCYTIC LEUKEMIA PATIENTS (CLL-IPI) AND COMPARISON WITH MDACC PROGNOSTIC INDEX: ANALYSIS OF 620 CASES
(Abstract release date: 05/19/16)
EHA Library. Gentile M. 06/09/16; 132611; E1062
Dr. Massimo Gentile
Contributions
Contributions
Abstract
Abstract: E1062
Type: Eposter Presentation
Background
In the last two decades, a plethora of clinical, serological and biological markers have been identified that are significantly associated with the prognosis of chronic lymphocytic leukemia (CLL) patients. Recent research has focused on the development of scoring systems capable of integrating the major prognostic parameters. A recent international collaboration proposed an international prognostic index (CLL-IPI) built on clinical, serological, and biological parameters (TP53 deletion and/or mutation, IGHV mutational status, β2M, clinical stage, and age) universally applicable to previously untreated CLL patients to predict overall survival (OS).
Aims
We performed a validation of the CLL-IPI in an independent series of Italian patients. Furthermore, we compared this tool with the prognostic index proposed by the MDACC CLL group in 2007.
Methods
Databases from 4 Italian centers including roughly 3000 newly diagnosed CLL patients were used to evaluate the validity and reproducibility of the CLL-IPI. Baseline data regarding age, Binet stage, IGHV mutational status, β2M and TP53 status, using only del17p, were available for 620 cases. The parameters for the calculation of the MDACC score were also available in these patients. The CLL-IPI and the MDACC score were calculated using the methods proposed. The accuracy of the prognostic models was assessed by the Harrell C index (an index of discrimination), the explained variation in mortality (an index combining calibration and discrimination), and the Akaike information criterion (AIC, an index comparing two non-nested prognostic models). The lower the AIC, the higher the prognostic accuracy of a predictive model.
Results
The median age of the 620 patients was 65 years (range 27-92) with 55.1% males. The majority of patients had Binet stage A (81.9%) and 366 cases (58.7%) had Rai stage 0. All 5 parameters of the CLL-IPI were found to be independently associated with survival in this study cohort (age >65 years: HR 5.48, P<0.0001; Binet B/C stage: HR 2.17, P<0.0001; β2M>3.5 mg/L: HR 1.99, P=0.001; IGHV unmutated: HR 1.82, P=0.002; TP53 deleted: HR 2.84, P<0.0001). According to the CLL-IPI, 57.1% of patients were classified as low-, 23.3% as intermediate-, 14.9% as high-, and 4.7% as very high-risk. The 5-year OS probabilities were: 93.5% for low-risk, 83.2% for intermediate-risk, 68.6% for high-risk, and 35.1% for very high-risk cases (P<0.0001; Harrell C index=73%; P<0.001). The prognostic index risk category remained a predictor of survival when analysis was limited to Rai stage 0 (P<0.0001).Subsequently, we compared the CLL-IPI with the MDACC prognostic index score in our cohort. The Harrell C index of the MDACC score was 70% (P<0.001), similar to the CLL-IPI (73%). The AIC showed the superiority of the CLL-IPI compared to the MDACC score in predicting OS (CLL-IPI, AIC=1372.007 versus MDACC score, AIC=1383.364). Accordingly, the explained variation in mortality provided by the CLL-IPI was 29% (P<0.001), a figure higher than that due to the MDACC score (26%, P<0.001), indicating that the CLL-IPI had a higher prognostic accuracy for mortality compared to that of MDACC score.
Conclusion
Our results confirm the validity of the CLL-IPI to predict survival among patients with previously untreated CLL. Moreover, we have demonstrated that the CLL-IPI which combines clinical and serological data with biological parameters has a higher accuracy for predicting prognosis of CLL patients than the MDACC score, a model built only on clinical and laboratory markers.
Session topic: E-poster
Keyword(s): Chronic lymphocytic leukemia, Prognosis
Type: Eposter Presentation
Background
In the last two decades, a plethora of clinical, serological and biological markers have been identified that are significantly associated with the prognosis of chronic lymphocytic leukemia (CLL) patients. Recent research has focused on the development of scoring systems capable of integrating the major prognostic parameters. A recent international collaboration proposed an international prognostic index (CLL-IPI) built on clinical, serological, and biological parameters (TP53 deletion and/or mutation, IGHV mutational status, β2M, clinical stage, and age) universally applicable to previously untreated CLL patients to predict overall survival (OS).
Aims
We performed a validation of the CLL-IPI in an independent series of Italian patients. Furthermore, we compared this tool with the prognostic index proposed by the MDACC CLL group in 2007.
Methods
Databases from 4 Italian centers including roughly 3000 newly diagnosed CLL patients were used to evaluate the validity and reproducibility of the CLL-IPI. Baseline data regarding age, Binet stage, IGHV mutational status, β2M and TP53 status, using only del17p, were available for 620 cases. The parameters for the calculation of the MDACC score were also available in these patients. The CLL-IPI and the MDACC score were calculated using the methods proposed. The accuracy of the prognostic models was assessed by the Harrell C index (an index of discrimination), the explained variation in mortality (an index combining calibration and discrimination), and the Akaike information criterion (AIC, an index comparing two non-nested prognostic models). The lower the AIC, the higher the prognostic accuracy of a predictive model.
Results
The median age of the 620 patients was 65 years (range 27-92) with 55.1% males. The majority of patients had Binet stage A (81.9%) and 366 cases (58.7%) had Rai stage 0. All 5 parameters of the CLL-IPI were found to be independently associated with survival in this study cohort (age >65 years: HR 5.48, P<0.0001; Binet B/C stage: HR 2.17, P<0.0001; β2M>3.5 mg/L: HR 1.99, P=0.001; IGHV unmutated: HR 1.82, P=0.002; TP53 deleted: HR 2.84, P<0.0001). According to the CLL-IPI, 57.1% of patients were classified as low-, 23.3% as intermediate-, 14.9% as high-, and 4.7% as very high-risk. The 5-year OS probabilities were: 93.5% for low-risk, 83.2% for intermediate-risk, 68.6% for high-risk, and 35.1% for very high-risk cases (P<0.0001; Harrell C index=73%; P<0.001). The prognostic index risk category remained a predictor of survival when analysis was limited to Rai stage 0 (P<0.0001).Subsequently, we compared the CLL-IPI with the MDACC prognostic index score in our cohort. The Harrell C index of the MDACC score was 70% (P<0.001), similar to the CLL-IPI (73%). The AIC showed the superiority of the CLL-IPI compared to the MDACC score in predicting OS (CLL-IPI, AIC=1372.007 versus MDACC score, AIC=1383.364). Accordingly, the explained variation in mortality provided by the CLL-IPI was 29% (P<0.001), a figure higher than that due to the MDACC score (26%, P<0.001), indicating that the CLL-IPI had a higher prognostic accuracy for mortality compared to that of MDACC score.
Conclusion
Our results confirm the validity of the CLL-IPI to predict survival among patients with previously untreated CLL. Moreover, we have demonstrated that the CLL-IPI which combines clinical and serological data with biological parameters has a higher accuracy for predicting prognosis of CLL patients than the MDACC score, a model built only on clinical and laboratory markers.
Session topic: E-poster
Keyword(s): Chronic lymphocytic leukemia, Prognosis
Abstract: E1062
Type: Eposter Presentation
Background
In the last two decades, a plethora of clinical, serological and biological markers have been identified that are significantly associated with the prognosis of chronic lymphocytic leukemia (CLL) patients. Recent research has focused on the development of scoring systems capable of integrating the major prognostic parameters. A recent international collaboration proposed an international prognostic index (CLL-IPI) built on clinical, serological, and biological parameters (TP53 deletion and/or mutation, IGHV mutational status, β2M, clinical stage, and age) universally applicable to previously untreated CLL patients to predict overall survival (OS).
Aims
We performed a validation of the CLL-IPI in an independent series of Italian patients. Furthermore, we compared this tool with the prognostic index proposed by the MDACC CLL group in 2007.
Methods
Databases from 4 Italian centers including roughly 3000 newly diagnosed CLL patients were used to evaluate the validity and reproducibility of the CLL-IPI. Baseline data regarding age, Binet stage, IGHV mutational status, β2M and TP53 status, using only del17p, were available for 620 cases. The parameters for the calculation of the MDACC score were also available in these patients. The CLL-IPI and the MDACC score were calculated using the methods proposed. The accuracy of the prognostic models was assessed by the Harrell C index (an index of discrimination), the explained variation in mortality (an index combining calibration and discrimination), and the Akaike information criterion (AIC, an index comparing two non-nested prognostic models). The lower the AIC, the higher the prognostic accuracy of a predictive model.
Results
The median age of the 620 patients was 65 years (range 27-92) with 55.1% males. The majority of patients had Binet stage A (81.9%) and 366 cases (58.7%) had Rai stage 0. All 5 parameters of the CLL-IPI were found to be independently associated with survival in this study cohort (age >65 years: HR 5.48, P<0.0001; Binet B/C stage: HR 2.17, P<0.0001; β2M>3.5 mg/L: HR 1.99, P=0.001; IGHV unmutated: HR 1.82, P=0.002; TP53 deleted: HR 2.84, P<0.0001). According to the CLL-IPI, 57.1% of patients were classified as low-, 23.3% as intermediate-, 14.9% as high-, and 4.7% as very high-risk. The 5-year OS probabilities were: 93.5% for low-risk, 83.2% for intermediate-risk, 68.6% for high-risk, and 35.1% for very high-risk cases (P<0.0001; Harrell C index=73%; P<0.001). The prognostic index risk category remained a predictor of survival when analysis was limited to Rai stage 0 (P<0.0001).Subsequently, we compared the CLL-IPI with the MDACC prognostic index score in our cohort. The Harrell C index of the MDACC score was 70% (P<0.001), similar to the CLL-IPI (73%). The AIC showed the superiority of the CLL-IPI compared to the MDACC score in predicting OS (CLL-IPI, AIC=1372.007 versus MDACC score, AIC=1383.364). Accordingly, the explained variation in mortality provided by the CLL-IPI was 29% (P<0.001), a figure higher than that due to the MDACC score (26%, P<0.001), indicating that the CLL-IPI had a higher prognostic accuracy for mortality compared to that of MDACC score.
Conclusion
Our results confirm the validity of the CLL-IPI to predict survival among patients with previously untreated CLL. Moreover, we have demonstrated that the CLL-IPI which combines clinical and serological data with biological parameters has a higher accuracy for predicting prognosis of CLL patients than the MDACC score, a model built only on clinical and laboratory markers.
Session topic: E-poster
Keyword(s): Chronic lymphocytic leukemia, Prognosis
Type: Eposter Presentation
Background
In the last two decades, a plethora of clinical, serological and biological markers have been identified that are significantly associated with the prognosis of chronic lymphocytic leukemia (CLL) patients. Recent research has focused on the development of scoring systems capable of integrating the major prognostic parameters. A recent international collaboration proposed an international prognostic index (CLL-IPI) built on clinical, serological, and biological parameters (TP53 deletion and/or mutation, IGHV mutational status, β2M, clinical stage, and age) universally applicable to previously untreated CLL patients to predict overall survival (OS).
Aims
We performed a validation of the CLL-IPI in an independent series of Italian patients. Furthermore, we compared this tool with the prognostic index proposed by the MDACC CLL group in 2007.
Methods
Databases from 4 Italian centers including roughly 3000 newly diagnosed CLL patients were used to evaluate the validity and reproducibility of the CLL-IPI. Baseline data regarding age, Binet stage, IGHV mutational status, β2M and TP53 status, using only del17p, were available for 620 cases. The parameters for the calculation of the MDACC score were also available in these patients. The CLL-IPI and the MDACC score were calculated using the methods proposed. The accuracy of the prognostic models was assessed by the Harrell C index (an index of discrimination), the explained variation in mortality (an index combining calibration and discrimination), and the Akaike information criterion (AIC, an index comparing two non-nested prognostic models). The lower the AIC, the higher the prognostic accuracy of a predictive model.
Results
The median age of the 620 patients was 65 years (range 27-92) with 55.1% males. The majority of patients had Binet stage A (81.9%) and 366 cases (58.7%) had Rai stage 0. All 5 parameters of the CLL-IPI were found to be independently associated with survival in this study cohort (age >65 years: HR 5.48, P<0.0001; Binet B/C stage: HR 2.17, P<0.0001; β2M>3.5 mg/L: HR 1.99, P=0.001; IGHV unmutated: HR 1.82, P=0.002; TP53 deleted: HR 2.84, P<0.0001). According to the CLL-IPI, 57.1% of patients were classified as low-, 23.3% as intermediate-, 14.9% as high-, and 4.7% as very high-risk. The 5-year OS probabilities were: 93.5% for low-risk, 83.2% for intermediate-risk, 68.6% for high-risk, and 35.1% for very high-risk cases (P<0.0001; Harrell C index=73%; P<0.001). The prognostic index risk category remained a predictor of survival when analysis was limited to Rai stage 0 (P<0.0001).Subsequently, we compared the CLL-IPI with the MDACC prognostic index score in our cohort. The Harrell C index of the MDACC score was 70% (P<0.001), similar to the CLL-IPI (73%). The AIC showed the superiority of the CLL-IPI compared to the MDACC score in predicting OS (CLL-IPI, AIC=1372.007 versus MDACC score, AIC=1383.364). Accordingly, the explained variation in mortality provided by the CLL-IPI was 29% (P<0.001), a figure higher than that due to the MDACC score (26%, P<0.001), indicating that the CLL-IPI had a higher prognostic accuracy for mortality compared to that of MDACC score.
Conclusion
Our results confirm the validity of the CLL-IPI to predict survival among patients with previously untreated CLL. Moreover, we have demonstrated that the CLL-IPI which combines clinical and serological data with biological parameters has a higher accuracy for predicting prognosis of CLL patients than the MDACC score, a model built only on clinical and laboratory markers.
Session topic: E-poster
Keyword(s): Chronic lymphocytic leukemia, Prognosis
{{ help_message }}
{{filter}}