scholarly journals The development and validation of prognostic models for overall survival in the presence of missing data in the training dataset: a strategy with a detailed example

2021 ◽  
Vol 5 (1) ◽  
Author(s):  
Kara-Louise Royle ◽  
David A. Cairns
Keyword(s):  
Overall Survival ◽  
Missing Data ◽  
Prognostic Model ◽  
Prognostic Index ◽  
Risk Groups ◽  
Prognostic Models ◽  
Training Dataset ◽  
Validation Process ◽  
Test Dataset ◽  
Development And Validation

Abstract Background The United Kingdom Myeloma Research Alliance (UK-MRA) Myeloma Risk Profile is a prognostic model for overall survival. It was trained and tested on clinical trial data, aiming to improve the stratification of transplant ineligible (TNE) patients with newly diagnosed multiple myeloma. Missing data is a common problem which affects the development and validation of prognostic models, where decisions on how to address missingness have implications on the choice of methodology. Methods Model building The training and test datasets were the TNE pathways from two large randomised multicentre, phase III clinical trials. Potential prognostic factors were identified by expert opinion. Missing data in the training dataset was imputed using multiple imputation by chained equations. Univariate analysis fitted Cox proportional hazards models in each imputed dataset with the estimates combined by Rubin’s rules. Multivariable analysis applied penalised Cox regression models, with a fixed penalty term across the imputed datasets. The estimates from each imputed dataset and bootstrap standard errors were combined by Rubin’s rules to define the prognostic model. Model assessment Calibration was assessed by visualising the observed and predicted probabilities across the imputed datasets. Discrimination was assessed by combining the prognostic separation D-statistic from each imputed dataset by Rubin’s rules. Model validation The D-statistic was applied in a bootstrap internal validation process in the training dataset and an external validation process in the test dataset, where acceptable performance was pre-specified. Development of risk groups Risk groups were defined using the tertiles of the combined prognostic index, obtained by combining the prognostic index from each imputed dataset by Rubin’s rules. Results The training dataset included 1852 patients, 1268 (68.47%) with complete case data. Ten imputed datasets were generated. Five hundred twenty patients were included in the test dataset. The D-statistic for the prognostic model was 0.840 (95% CI 0.716–0.964) in the training dataset and 0.654 (95% CI 0.497–0.811) in the test dataset and the corrected D-Statistic was 0.801. Conclusion The decision to impute missing covariate data in the training dataset influenced the methods implemented to train and test the model. To extend current literature and aid future researchers, we have presented a detailed example of one approach. Whilst our example is not without limitations, a benefit is that all of the patient information available in the training dataset was utilised to develop the model. Trial registration Both trials were registered; Myeloma IX-ISRCTN68454111, registered 21 September 2000. Myeloma XI-ISRCTN49407852, registered 24 June 2009.

A Comparison of Prognostic Models for Chronic Myelomonocytic Leukemia (CMML) in the Era of Hypomethylating Agents

Blood ◽  
2012 ◽  
Vol 120 (21) ◽  
pp. 1695-1695 ◽  
Author(s):  
Eric Padron ◽  
Najla H Al Ali ◽  
Deniz Peker ◽  
Jeffrey E Lancet ◽  
Pearlie K Epling-Burnette ◽  
...  
Keyword(s):  
Overall Survival ◽  
Scoring System ◽  
Prognostic Model ◽  
Median Overall Survival ◽  
Prognostic Score ◽  
Risk Groups ◽  
Prognostic Models ◽  
Rank Test ◽  
Log Rank Test ◽  
Md Anderson

Abstract Abstract 1695 Introduction: CMML is a genetically and clinically heterogeneous malignancy characterized by peripheral monocytosis, cytopenias, and a propensity for AML transformation. Several prognostic models attempt to stratify patients into subcategories that are predictive for overall survival (OS), six models of which are specific to CMML. However, these models have either never been externally validated in the context of CMML or were externally validated prior to the use of hypomethylating agents. We externally validate and perform a detailed statistical comparison between the International Prognostic Scoring System (IPSS), MD Anderson Scoring System (MDASC), MD Anderson Prognostic Score (MDAPS), Dusseldorf Score (DS), and Spanish Scoring Systems (SS) in a large, single institution cohort. Methods: Data were collected retrospectively from the Moffitt Cancer Center (MCC) CMML database and charts were reviewed of patients that satisfied the WHO criteria for the diagnosis of CMML. The primary objective of the study was to validate the above prognostic models calculated at the time of initial presentation to MCC. All prognostic models were calculated as previously published. All analyses were conducted using SPSS version 15.0 (SPSS Inc, Chicago, IL). The Kaplan–Meier (KM) method was used to estimate median overall survival and the log rank test was used to compare KM survival estimates between two groups. Results: Between January 2000 and February 2012, 123 patients were captured by the MCC CMML database. The median age at diagnosis was 69 (30–90) years and the majority of patients were male (69%). By the WHO classification, the majority of patients had CMML-1 (84% vs. 16%) and most patients were subcategorized as MPN-CMML (59%) versus MDS-CMML (39%) by the FAB CMML criteria. The median overall survival of the entire cohort was 30 months and the rate of AML transformation was 44% (54). Twenty-two patients (18%) were treated with decitabine and 66 (54%) patients were treated with 5-azacitidine. Risk group stratification according to specific prognostic model is summarized in Table 1. The IPSS, MDASC, DS, and SS all predicted OS (p<0.05) while the MDASP could not be validated (p=0.924). When only patients who were treated with 5-azacitadine were considered, the MDASC, DS, and SS continued to predict OS (p<0.05) while the IPSS (p=0.15) and MDASP (p=0.239) did not. Previous reports have demonstrated that the MDASC provides further discrimination to refine stratification by the IPSS in Myelodysplastic Syndromes (MDS). Except for the low-risk DS patients, we grouped patients in our CMML cohort into lower and higher risk disease with each prognostic score and attempted to further stratify patients by the MDASC using KM and the log rank test. The MDASC was able to further risk stratify patients in each group for all prognostic models except those in the higher risk groups by the SS (p=0.07) and DS (P=0.45). When a similar statistical analysis was applied to each prognostic scoring system, only the MDASC was consistently able to further stratify the majority of risk groups as described in Table 2. The Dusseldorf scoring system was able to further stratify all lower risk groups regardless of model but was not able to do so in higher risk disease. Conclusions: This represents the first external validation of existing CMML prognostic models in the era of hypomethylating agent therapy. Except for the MDASP, we were able to validate the prognostic value all models tested. The MDASC represents the most robust model as it consistently refined the stratification of other models tested and remained predictive of OS in 5-azacitidine treated patients. Multi-institution collaboration is needed to construct a robust CMML specific prognostic model. Comparison to the IPSS-R is in progress. Disclosures: No relevant conflicts of interest to declare.

A New Prognostic Model for Extranodal Natural Killer/T Cell Lymphoma, Nasal Type

Blood ◽  
2013 ◽  
Vol 122 (21) ◽  
pp. 1769-1769
Author(s):  
Qingqing Cai ◽  
Xiaolin Luo ◽  
Ken H. Young ◽  
Huiqiang Huang ◽  
Guanrong Zhang ◽  
...  
Keyword(s):  
Overall Survival ◽  
Prognostic Factors ◽  
High Risk ◽  
Prognostic Model ◽  
Cell Lymphoma ◽  
Prognostic Index ◽  
Roc Curves ◽  
Risk Groups ◽  
Survival Outcomes ◽  
Nasal Type

Abstract Background Extranodal natural killer (NK)/T–cell lymphoma, nasal type (ENKTL) is an aggressive disease with a poor prognosis. A better risk stratification is beneficial for clinical management in affected patients. Our recent study has shown that fasting blood glucose (FBG) was a novel, prognostic factor, (Cai et al, British Journal of Cancer, 108: 380–386,2013). This finding has not been integrated in the previous prognostic models for ENKTL Therefore, we aimed to design a new prognostic model, including FBG, for ENKTL which supports to identify high–risk patients eligible for advanced or more aggressive therapy. Patients and methods 158 newly diagnosed patients with ENKTL were analyzed between January 2003 and January 2011 at Sun Yat–sen University Cancer Center, China. Overall survival (OS) and progression free survival (PFS) were estimated using the Kaplan–Meier method. The significance of differences between survival was tested using the Log–rank test. Significant variables in the univariate analysis were selected as variables for the multivariate analysis of survival. The latter was performed by the Cox regression mode. We constructed receiver operating characteristic (ROC) curves and compared the areas under the ROC curves of total protein (TP), FBG, Korean Prognostic Index (KPI) and their combinations in comparison to the survival outcome. Results Of 158 patients, 156 patients had complete clinical information for the parameters of the International Prognostic Index (IPI) model and KPI model. The estimated 5–year overall survival rate in 158 patients was 59.2%. Independent prognostic factors included TP < 60 g/L, FBG > 100 mg/dL, KPI score ≥ 2. A new prognostic model was constructed by combining these prognostic factors: Group 1 (64 cases, 41.0%), no adverse factors; Group 2 (58 cases, 37.2%), one adverse factor; and Group 3 (34 cases, 21.8%), two or three adverse factors. The 5–year overall survival of these groups were 88.9%, 35.6% and 12.7%, respectively (p < 0.001). The survival curves according to the new prognostic model are shown in Fig. 1. The new model categorized three groups with significantly different survival outcomes. The new prognostic model was also efficient in discriminating the patients with low to low–intermediate risk IPI group and high–intermediate to high risk IPI group into three subgroups with different survival outcomes (p < 0.001). The KPI model balanced the distribution of patients into different risk groups better than IPI prognostic model (score 0: 12 cases, 7.7%; score 1: 38 cases, 24.4%; score 2: 42 cases, 26.9%; score 3–4: 64 cases, 41.0%), and it was able to differentiate patients with different survival outcomes (p < 0.001). In addition, the new prognostic model had a better prognostic value than did KPI model alone (p < 0.001), suggesting that TP and FBG reinforced the prognostic ability of KPI model (Table 1). Conclusions The new prognostic model we proposed for ENKTL, including the new prognostic indicator total protein and FBG, demonstrated balanced distribution of patients into different risk groups with better prognostic discrimination as compared to KPI model alone. Disclosures: No relevant conflicts of interest to declare.

Prognostic model for survival in patients with advanced pancreatic cancer receiving palliative chemotherapy.

2015 ◽  
Vol 33 (3_suppl) ◽  
pp. 248-248
Author(s):  
Yu Uneno ◽  
Tadayuki Kou ◽  
Masashi Kanai ◽  
Michio Yamamoto ◽  
Peng Xue ◽  
...  
Keyword(s):  
Pancreatic Cancer ◽  
Overall Survival ◽  
Prognostic Factors ◽  
Poor Prognosis ◽  
Prognostic Model ◽  
Palliative Chemotherapy ◽  
Cox Regression ◽  
Advanced Pancreatic Cancer ◽  
Prognostic Index ◽  
Prognostic Models

248 Background: The prognosis of patients with advanced pancreatic cancer (APC) is extremely poor. Several clinical and laboratory factors have been known to be associated with prognosis of APC patients. However, there are few clinically available prognostic models predicting survival in APC patients receiving palliative chemotherapy. Methods: To construct a prognostic model to predict survival in APC patients receiving palliative chemotherapy, we analyzed the clinical data from 306 consecutive patients with pathologically confirmed APC who received palliative chemotherapy. We selected six independent prognostic factors which remained independent prognostic factors after multivariate analysis. Thereafter, we rounded the regression coefficient (β) for each independent prognostic factor derived from the Cox regression equation (HR = eβ) and developed a prognostic index (PI). Results: Developed prognostic index (PI) was as follows: PI = 2 (if performance status score 2–3) + 1 (if metastatic disease) + 1 (if initially unresectable disease) + 1 (if carcinoembryonic antigen level ≥5.0 ng/ml) + 1 (if carbohydrate antigen 19-9 level ≥1000 U/ml) + 2 (if neutrophil–lymphocyte ratio ≥5). The patients were classified into three prognostic groups: favorable (PI 0–1, n = 73), intermediate (PI 2–3, n = 145), and poor prognosis (PI 4–8, n = 88). The median overall survival for each prognostic group was 16.5, 12.3 and 6.2 months, respectively, and the 1-year survival rates were 67.3%, 51.3%, and 19.1%, respectively (P < 0.01). The c index of the model was 0.658. This model was well calibrated to predict 1-year survival, in which overestimation (2.4% and 0.2% in the favorable and poor prognosis groups, respectively) and underestimation (3.6% in the intermediate prognosis group) were observed. Conclusions: This prognostic model based on readily available clinical factors would help clinicians in estimating the overall survival in APC patients receiving palliative chemotherapy.

scholarly journals Development and validation of an individualized immune prognostic model in stage I–III lung squamous cell carcinoma

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Qi-Fan Yang ◽  
Di Wu ◽  
Jian Wang ◽  
Li Ba ◽  
Chen Tian ◽  
...  
Keyword(s):  
Overall Survival ◽  
Squamous Cell Carcinoma ◽  
Cell Carcinoma ◽  
Squamous Cell ◽  
Prognostic Model ◽  
Lung Squamous Cell Carcinoma ◽  
Risk Groups ◽  
Stage I ◽  
Tissue Samples ◽  
Mutation Burden

AbstractLung squamous cell carcinoma (LUSC) possesses a poor prognosis even for stages I–III resected patients. Reliable prognostic biomarkers that can stratify and predict clinical outcomes for stage I–III resected LUSC patients are urgently needed. Based on gene expression of LUSC tissue samples from five public datasets, consisting of 687 cases, we developed an immune-related prognostic model (IPM) according to immune genes from ImmPort database. Then, we comprehensively analyzed the immune microenvironment and mutation burden that are significantly associated with this model. According to the IPM, patients were stratified into high- and low-risk groups with markedly distinct survival benefits. We found that patients with high immune risk possessed a higher proportion of immunosuppressive cells such as macrophages M0, and presented higher expression of CD47, CD73, SIRPA, and TIM-3. Moreover, When further stratified based on the tumor mutation burden (TMB) and risk score, patients with high TMB and low immune risk had a remarkable prolonged overall survival compared to patients with low TMB and high immune risk. Finally, a nomogram combing the IPM with clinical factors was established to provide a more precise evaluation of prognosis. The proposed immune relevant model is a promising biomarker for predicting overall survival in stage I–III LUSC. Thus, it may shed light on identifying patient subset at high risk of adverse prognosis from an immunological perspective.

scholarly journals A prognostic model based on seven immune-related genes predicts the overall survival of patients with hepatocellular carcinoma

2021 ◽  
Vol 14 (1) ◽  
Author(s):  
Qian Yan ◽  
Wenjiang Zheng ◽  
Boqing Wang ◽  
Baoqian Ye ◽  
Huiyan Luo ◽  
...  
Keyword(s):  
Hepatocellular Carcinoma ◽  
Overall Survival ◽  
High Risk ◽  
Risk Score ◽  
Immune Cells ◽  
Prognostic Model ◽  
Risk Groups ◽  
Low Risk ◽  
Prognostic Performance ◽  
Immune Related Genes

Abstract Background Hepatocellular carcinoma (HCC) is a disease with a high incidence and a poor prognosis. Growing amounts of evidence have shown that the immune system plays a critical role in the biological processes of HCC such as progression, recurrence, and metastasis, and some have discussed using it as a weapon against a variety of cancers. However, the impact of immune-related genes (IRGs) on the prognosis of HCC remains unclear. Methods Based on The Cancer Gene Atlas (TCGA) and Immunology Database and Analysis Portal (ImmPort) datasets, we integrated the ribonucleic acid (RNA) sequencing profiles of 424 HCC patients with IRGs to calculate immune-related differentially expressed genes (DEGs). Survival analysis was used to establish a prognostic model of survival- and immune-related DEGs. Based on genomic and clinicopathological data, we constructed a nomogram to predict the prognosis of HCC patients. Gene set enrichment analysis further clarified the signalling pathways of the high-risk and low-risk groups constructed based on the IRGs in HCC. Next, we evaluated the correlation between the risk score and the infiltration of immune cells, and finally, we validated the prognostic performance of this model in the GSE14520 dataset. Results A total of 100 immune-related DEGs were significantly associated with the clinical outcomes of patients with HCC. We performed univariate and multivariate least absolute shrinkage and selection operator (Lasso) regression analyses on these genes to construct a prognostic model of seven IRGs (Fatty Acid Binding Protein 6 (FABP6), Microtubule-Associated Protein Tau (MAPT), Baculoviral IAP Repeat Containing 5 (BIRC5), Plexin-A1 (PLXNA1), Secreted Phosphoprotein 1 (SPP1), Stanniocalcin 2 (STC2) and Chondroitin Sulfate Proteoglycan 5 (CSPG5)), which showed better prognostic performance than the tumour/node/metastasis (TNM) staging system. Moreover, we constructed a regulatory network related to transcription factors (TFs) that further unravelled the regulatory mechanisms of these genes. According to the median value of the risk score, the entire TCGA cohort was divided into high-risk and low-risk groups, and the low-risk group had a better overall survival (OS) rate. To predict the OS rate of HCC, we established a gene- and clinical factor-related nomogram. The receiver operating characteristic (ROC) curve, concordance index (C-index) and calibration curve showed that this model had moderate accuracy. The correlation analysis between the risk score and the infiltration of six common types of immune cells showed that the model could reflect the state of the immune microenvironment in HCC tumours. Conclusion Our IRG prognostic model was shown to have value in the monitoring, treatment, and prognostic assessment of HCC patients and could be used as a survival prediction tool in the near future.

Mantle Cell Lymphoma: Prognostic Capacity of the Follicular Lymphoma International Prognostic Index.

Blood ◽  
2005 ◽  
Vol 106 (11) ◽  
pp. 1910-1910
Author(s):  
Michael B. Moller ◽  
Niels T. Pedersen ◽  
Bjarne E. Christensen
Keyword(s):  
Overall Survival ◽  
Multivariate Analysis ◽  
Mantle Cell Lymphoma ◽  
Cell Lymphoma ◽  
International Prognostic Index ◽  
Prognostic Index ◽  
Risk Groups ◽  
High Risk Group ◽  
Predictive Capacity ◽  
Mantle Cell

Abstract Background: The International Prognostic Index (IPI) is the most commonly used prognostic model in mantle cell lymphoma. However, the prognostic value of IPI is limited. The recently published Follicular Lymphoma International Prognostic Index (FLIPI) is built on variables (age, stage, lactic dehydrogenase, anemia, and nodal disease) which also are pertinent to mantle cell lymphoma. This study was conducted to evaluate the prognostic value of FLIPI in patients with mantle cell lymphoma. Patients and Methods: A population-based series of 93 patients with mantle cell lymphoma diagnosed in a 7-year period were studied. End points of the study were response to therapy, overall survival, and failure-free survival according to IPI and FLIPI. Results: Applied to the whole series, FLIPI identified 3 risk groups with markedly different outcome with 5-year overall survival rates of 65%, 42%, and 8%, respectively (P < .0001; log-rank 28.13; figure below). Notably, the high-risk group comprised 53% of patients. In contrast, IPI only allocated 16% of cases to the high-risk group and had a lower overall predictive capacity (log-rank 24.8). When both FLIPI and IPI were included in a multivariate analysis, only FLIPI was related to survival. In patients treated with CHOP-based regimens (n = 45) FLIPI also had superior predictive capacity compared to IPI (log-rank, 18.51 versus 11.37), and again only FLIPI retained significance in multivariate analysis. Multivariate analysis of failure-free survival also identified FLIPI, and not IPI, as independently significant. Conclusion: FLIPI is the superior prognostic model as compared to IPI and should be the preferred clinical prognostic index in mantle cell lymphoma. Overall survival according to FLIPI risk groups Overall survival according to FLIPI risk groups

An Enhanced International Prognostic Index (IPI) for Patients with Diffuse Large B-Cell Lymphoma (DLBCL) in the Rituximab Era Using the National Comprehensive Cancer Network (NCCN) Database.

Blood ◽  
2012 ◽  
Vol 120 (21) ◽  
pp. 2656-2656
Author(s):  
Zheng Zhou ◽  
Alfred W. Rademaker ◽  
Leo I. Gordon ◽  
Ann S. LaCasce ◽  
Ann Vanderplas ◽  
...  
Keyword(s):  
Bone Marrow ◽  
Overall Survival ◽  
High Risk ◽  
International Prognostic Index ◽  
Model Development ◽  
Prognostic Index ◽  
Risk Groups ◽  
Ratio Test ◽  
Validation Sample ◽  
Gi Tract

Abstract Abstract 2656 Introduction: The International Prognostic Index (IPI) was first developed in 1993 to risk stratify patients with aggressive Non-Hodgkin's lymphoma (NHL) for outcome prediction (Shipp, NEJM, 1993). Since the addition of rituximab to conventional CHOP chemotherapy for the treatment of DLBCL, there have been many efforts to validate the IPI as well as to improve upon the model's capacity to distinguish subgroups with discrete clinical outcomes, especially high-risk patients. Previous studies have focused on adding clinical or biologic prognostic factor(s) to the original model or regrouping the original IPI score (R-IPI; Sehn, Blood, 2007). We, therefore, built anew a modern IPI based solely on clinical factors available in the real world NCCN clinical database. Methods: Using the nationwide population-based NHL lymphoma database from NCCN, patients with newly diagnosed DLBCL enrolled between June 2000 and Dec. 2010 at 7 NCCN cancer centers were included with at least 1 year and up to 5 years of follow-up. Clinical characteristics including age, Ann Arbor stage, ECOG performance status, disease in extranodal sites (including positivity in bone marrow, CNS, liver/GI tract, lung, other sites and spleen), LDH, presence of bulky disease (>10 cm) as well as B symptoms were studied as potential predictors for model development using COX proportional hazards regression. IPI scores were assigned proportionally based on parameter estimates of the statistically significant predictors in the final COX model. Model selection and its comparison to the original IPI model were made based on Akaike Criteria (AIC) and the likelihood ratio test. Categorization of age and LDH was decided by testing the linearity assumption and Martingale residuals. Kaplan-Meier curves were plotted for stratified risk groups per the new and original IPI. Finally, both IPI models were compared using the initial randomly selected 15% validation sample. Results: There were 1,650 DLBCL patients with complete clinical information included for model development. The new IPI model consisted of similar component predictors but used a maximum of 8 scoring points by further categorizing age group into >40–60 (score of 1), >60–75 (score of 2) and >75 yrs (score of 3), and normalized LDH between >1–3 times (score of 1) and 33 times (score of 2) upper limit of normal. These categorizations minimized the Martingale residuals. Age effect was linear and 20-year increments fit the model best, whereas the effect of normalized LDH was not linear and reached plateau at a ratio of 3. Lymphomatous involvement either of bone marrow, CNS, Liver/GI tract or lung appeared as a stronger predictor (p<0.001) than number of extranodal sites (p=0.91). Four risk groups (Low, Low-intermediate, High-intermediate and High) were identified using the current IPI (Table 1) with enhanced discrimination power when compared with the original IPI and better global model fitting statistics, i.e. smaller AIC and significant likelihood ratio test (p<0.001). It was possible to identify a high risk group (score 3 6) with 5-year overall survival of 33% (95% CI: 22%–45%). Better model prediction was also shown in the validation sample. Conclusions: We were able to develop an enhanced IPI model for clinical prediction among previously untreated DLBCL cases by using patient level data from the NCCN NHL database. The NCCN-IPI demonstrates better risk stratification and identifies a poor risk subgroup with <50% 5-year overall survival in the current real-world clinical setting as compared to the original IPI model developed for aggressive lymphoma prior to the rituximab era. Disclosures: No relevant conflicts of interest to declare.

Treatment Outcome and Prognostic Model for Stage IE/IIE Extranodal Natural Killer/T-Cell Lymphoma, Nasal Type: A Retrospective Study in a Cohort of 305 Patients with Long Term Follow-up

Blood ◽  
2015 ◽  
Vol 126 (23) ◽  
pp. 2718-2718
Author(s):  
Yuankai Shi ◽  
Bo Jia ◽  
Xiaohui He ◽  
Youwu Shi ◽  
Mei Dong ◽  
...  
Keyword(s):  
Prognostic Factors ◽  
T Cell ◽  
Prognostic Model ◽  
Prognostic Index ◽  
Risk Groups ◽  
Natural Killer T Cell ◽  
Combined Chemoradiotherapy ◽  
Radiotherapy Alone ◽  
Killer T Cell

Abstract Background Extranodal natural killer/T-cell lymphoma, nasal type (ENKL) is a rare and distinct subtype of non-hodgkin lymphoma (NHL). The frequency was higher in Asia than in western countries and it has become the most common subtype of peripheral T-cell lymphomas in China. The majority of ENKL patients present with early stage. Optimal treatment modalities and prognostic factors for localized ENKL have not been fully defined. This study aimed to evaluate the optimal treatment strategy and prognostic factors for localized ENKL patients. Methods Between 2003 and 2013, three hundred and five patients with stage IE/IIE ENKL were comprehensively analyzed in this study. A total of 180 patients received combined chemoradiotherapy, with 111 patients received radiotherapy alone and 14 patients recieved chemotherapy alone. Chemotherapy regimens include GDP (gemcitabine, cisplatin, and dexamethasone), CHOP (epirubicin, cyclophosphamide, vincristine, and prednisolone) and other regimens. A total dose of 50 Gy to the primary tumor was considered as radical dose for ENKL, and additional 5 to 10 Gy was administered as a boost to the residual disease. Results The complete response (CR) rate for patients received chemoradiotherapy (n=175) was significantly higher than that for patients received radiotherapy alone (n=102) (89.1 % vs.77.5 %, P = 0.009) or chemotherapy alone (n=14) (89.1 % vs.21.4 %, P< 0.001). The median follow up time for all 305 patients was 38.7 (1.1 to 393) months. For 228 stage IE paranasal extension or IIE patients, 3-year overall survival (OS) in combined chemoradiotherapy (n=154), radiotherapy alone (n=60) and chemotherapy alone (n=14) groups were 85.7%, 73.3% and 57.1% respectively (chemoradiotherapy vs. radiotherapy, P=0.003; chemoradiotherapy vs. chemotherapy, P<0.001). For patients received combined chemoradiotherapy, GDP regimen (n=54) (included 10 patients with pegaspargase) could significantly improve 3-year progression-free survival (PFS) compared with CHOP-like (n=110) (included 10 patients with asparaginase) (88.9% vs. 70.9%, P =0.015).Patients received radiotherapy first followed by chemotherapy (n=84) was associated with superior 3-year PFS compared with patients initially received chemotherapy (n=96) (81.0% vs. 69.8%, P=0.034). But for 54 patients received GDP regimen, induction chemotherapy (n=17) could increase 3-year PFS (100.0% vs. 83.8%, P=0.112) and OS (100.0% vs. 86.5%, P=0.180). We identified 3 risk groups based on 3 prognostic factors (stage II, LDH elevated and paranasal extension) with different survival outcomes. The 3-year OS rates were 93.5%, 85.0% and 62.2% respectively for patients with no risk factors, 1 or 2 factors and 3 factors (P<0.001). Conclusions Combined chemoradiotherapy is the most optimal therapy strategy for stage IE paranasal extension or IIE ENKL patients. GDP or combined with pegaspargase regimen shows promising efficacy, significant superior to the traditional CHOP regimen. The sequence of chemotherapy and radiotherapy for patients received novel chemotherapy regimens still needs further assessment in phase 3 clinical trials. We identified 3 risk groups based on 3 prognostic factors (stage II, LDH elevated and paranasal extension) with different survival outcomes and this novel prognostic model may better predict prognosis than previous International Prognostic Index (IPI) and Korean Prognostic Index (KPI) score for ENKL patients with limited stage. Disclosures No relevant conflicts of interest to declare.

Identifying patients with non-small cell lung cancer (NSCLC) unlikely to benefit from erlotinib: An exploratory analysis of National Cancer of Institute of Canada Clinical Trials Group BR.21

2006 ◽  
Vol 24 (18_suppl) ◽  
pp. 7161-7161 ◽  
Author(s):  
M. Florescu ◽  
B. Hasan ◽  
F. A. Shepherd ◽  
L. Seymour ◽  
K. Ding ◽  
...  
Keyword(s):  
Weight Loss ◽  
Prognostic Model ◽  
Cox Regression ◽  
Prognostic Index ◽  
Early Death ◽  
Risk Groups ◽  
Smoking History ◽  
Factors Associated ◽  
Significant Survival ◽  
High Benefit

7161 Background: Despite a 9% response rate, BR.21 demonstrated significant survival benefit for patients receiving erlotinib as 2nd/3rd line therapy for NSCLC. We undertook to characterize, by exploratory subset analysis, patients less likely to benefit from erlotinib. To identify factors for consideration, we first identified baseline characteristics associated with early progression by eight wks and early death by 3 mos. Methods: Using stratification factors and potential prognostic factors from BR.21, the Cox regression model with stepwise selection was used to establish a prognostic model to separate erlotinib patients into 4 risk categories based on the 10th, 50th & 90th percentiles of prognostic index scores. 7 variables (smoking history, PS, weight loss, anemia, high LDH, response to prior chemo and time from diagnosis to randomization) were used in the final model. The hypothesis was that the characteristics of the treated patients in the highest risk group would also be predictive of lack of benefit from erlotinib when erlotinib and placebo patients with the same characteristics were compared. Results: Factors associated with PD by 8 wks were: PS2–3 (p = 0.009), weight loss (p = 0.0004), anemia (p = 0.008), PD to prior chemo (p = 0.006), non-Asian (p = 0.047), EGFR IHC-negative (p = 0.005), Factors associated with survival < 3 mos were: PS2–3 (p < 0.0001), weight loss (p < 0.0001), anemia (p < 0.0001), PD to prior chemo (p < 0.0001), non-Asian (p = 0.008), high LDH (p < 0.0001), time to randomization <12 mos (p = 0.0003). Comparison of overall survival for the 4 risk groups derived from prognostic index score as follows: high benefit (HR = 0.41, p = 0.007), 2 intermediate benefit (HR 0.79, p = 0.09; HR 0.80; p = 0.09); no benefit (HR 1.23; p = 0.42). Median survivals for erlotinib (placebo) patients in each group were 17.3 (8.3), 9.7 (7.5), 4.1 (3.7), 1.9 (2.7) mos. Conclusions: By establishing a prognostic model, we identified a small group of patients who are unlikely to benefit from 2nd/3rd line erlotinib therapy. This model requires prospective validation to confirm that it is both prognostic and predictive of outcome from treatment. [Table: see text]

Development and validation of a prognostic model for overall survival in chemotherapy-naive men with metastatic castration-resistant prostate cancer (mCRPC) from the phase 3 prevail clinical trial.

2017 ◽  
Vol 35 (15_suppl) ◽  
pp. 5022-5022
Author(s):  
Andrew J. Armstrong ◽  
Ping Lin ◽  
Celestia S. Higano ◽  
Cora N. Sternberg ◽  
Guru Sonpavde ◽  
...  
Keyword(s):  
Clinical Trial ◽  
Overall Survival ◽  
Prognostic Factors ◽  
High Risk ◽  
Prognostic Model ◽  
Risk Groups ◽  
Multivariable Model ◽  
Training Set ◽  
Phase 3 ◽  
Testing Set

5022 Background: Prognostic models require updating to reflect contemporary medical practice. In a post hoc analysis of the phase 3 PREVAIL trial (enzalutamide vs placebo), we identified prognostic factors for overall survival (OS) in chemotherapy-naive men with mCRPC. Methods: Patients were randomly divided 2:1 into training (n = 1159) and testing (n = 550) sets. Using the training set, 23 predefined candidate prognostic factors (including treatment) were analyzed in a multivariable Cox model with stepwise procedures and in a penalized Cox proportional hazards model using the adaptive least absolute shrinkage and selection operator (LASSO) penalty (data cutoff June 1, 2014). A multivariable model predicting OS was developed using the training set; the predictive accuracy was assessed in the testing set using time-dependent area under the curve (tAUC). The testing set was stratified based on risk score tertiles (low, intermediate, high), and OS was analyzed using Kaplan-Meier methodology. Results: Demographics, disease characteristics, and OS were balanced between the training and testing sets; median OS was 32.7 months for both datasets. There were no enzalutamide treatment-prognostic factor interactions (predictors). The final multivariable model included 11 prognostic factors: prostate-specific antigen, treatment, hemoglobin, neutrophil-lymphocyte ratio, liver metastases, time from diagnosis to randomization, lactate dehydrogenase, ≥ 10 bone metastases, pain, albumin, and alkaline phosphatase. The tAUC was 0.74 in the testing set. Median (95% confidence interval [CI]) OS for the low-, intermediate-, and high-risk groups (testing set) were not yet reached (NYR) (NYR–NYR), 34.2 months (31.5–NYR), and 21.1 months (17.5–25.0). The hazard ratios (95% CI) for OS in the low- and intermediate-risk groups vs the high-risk group were 0.20 (0.14–0.29) and 0.40 (0.30–0.53), respectively. Conclusions: Our validated prognostic model incorporates factors routinely collected in chemotherapy-naive men with mCRPC treated with enzalutamide and identifies subsets of men with widely differing survival times. Clinical trial information: NCT01212991.

Sign in / Sign up

Export Citation Format

Share Document