Harrell's Concordance Index Research Articles

Development and Validation of Nomograms Based on Nutritional Risk Index for Predicting Extracapsular Extension and Seminal Vesicle Invasion in Patients Undergoing Radical Prostatectomy.

The aim of the study was to investigate the predictive value of the nutritional risk index (NRI) for extracapsular extension (ECE) and seminal vesicle invasion (SVI) in prostate cancer (PCa) patients undergoing radical prostatectomy (RP), and further develop and validate predictive nomograms for ECE and SVI based on the NRI. We retrospectively analyzed 734 PCa patients who underwent RP between 2010 and 2020 in the Department of Urology at Peking University Third Hospital. The enrolled patients were randomly divided into a primary cohort (n = 489) and a validation cohort (n = 245) in a 2:1 manner. The baseline NRI of patients was calculated using serum albumin level and body mass index, and a malnutrition status was defined as NRI ≤ 98. Univariate and multivariate logistic regression analyses were conducted to identify predictors for ECE and SVI. Nomograms for predicting ECE and SVI were established based on the results of the multivariate logistic regression analysis. The performance of the nomograms was estimated using Harrell's concordance index (C-index), the area under curve (AUC) of receiver operating characteristic (ROC) curves and the calibration curves. In the primary cohort, 70 (14.3%) patients with NRI ≤ 98 were classified as malnutrition, while the remaining 419 (85.7%) patients with NRI > 98 were considered to have normal nutrition. The nomograms for predicting ECE and SVI shared common factors including NRI, percentage of positive biopsy cores (PPC) and biopsy Gleason score, while prostate-specific antigen (PSA) levels and PSA density (PSAD) were only incorporated in ECE nomogram. The C-indexes of the nomograms for predicting ECE and SVI were 0.785 (95% confidence interval (CI): 0.745 - 0.826) and 0.852 (95% CI: 0.806 - 0.898), respectively. The calibration curves demonstrated excellent agreement between the predictions by the nomograms and the actual observations. The results remained reproducible when the nomograms were applied to the validation cohort. The NRI is significantly associated with ECE and SVI in PCa patients. The nomogram established based on the NRI in our study can provide individualized risk estimation for ECE and SVI in PCa patients, and may be valuable for clinicians in making well-informed decisions regarding treatment strategies and patient management.

A risk signature based on endoplasmic reticulum stress-associated genes predicts prognosis and immunity in pancreatic cancer.

Introduction: The involvement of endoplasmic reticulum (ER) stress in cancer biology is increasingly recognized, yet its role in pancreatic cancer (PC) remains unclear. This study aims to elucidate the impact of ER stress on prognosis and biological characteristics in PC patients. Methods: A bioinformatic analysis was conducted using RNA-seq data and clinicopathological information from PC patients in the TCGA and ICGC databases. The ER stress-associated gene sets were extracted from MSigDB. ER stress-associated genes closely linked with overall survival (OS) of PC patients were identified via log-rank test and univariate Cox analysis, and further narrowed by LASSO method. A risk signature associated with ER stress was formulated using multivariate Cox regression and assessed through Kaplan-Meier curves, receiver operating characteristic (ROC) analyses, and Harrell's concordance index. External validation was performed with the ICGC cohort. The single-sample gene-set enrichment analysis (ssGSEA) algorithm appraised the immune cell infiltration landscape. Results: Worse OS in PC patients with high-risk signature score was observed. Multivariate analysis underscored our ER stress-associated signature as a valuable and independent predictor of prognosis. Importantly, these results based on TCGA were further validated in ICGC dataset. In addition, our risk signature was closely associated with homeostasis, protein secretion, and immune regulation in PC patients. In particular, PC microenvironment in the high-risk cluster exhibited a more immunosuppressive status. At last, we established a nomogram model by incorporating the risk signature and clinicopathological parameters, which behaves better in predicting prognosis of PC patients. Discussion: This comprehensive molecular analysis presents a new predictive model for the prognosis of PC patients, highlighting ER stress as a potential therapeutic target. Besides, the findings indicate that ER stress can have effect modulating PC immune responses.

Combining Gene Mutation with Transcriptomic Data Improves Outcome Prediction in Myelodysplastic Syndromes

Background and Aim. Myelodysplastic syndromes (MDS) are myeloid neoplasms characterized by peripheral blood cytopenias and risk of progression to acute myeloid leukemia (AML). Disease management is challenged by heterogeneity in clinical courses and survival probability. Recently, the genomic screening integration (by Molecular International Prognostic Scoring System, IPSS-M) into patient's assessment has resulted into a significant improvement in predicting clinical outcomes compared to the conventional prognostic score (Revised IPSS, IPSS-R). Many of the consequences of genetic and cytogenetic alterations will affect gene expression by means of transcriptional and epigenetic instability and altered microenviromental signaling. The aim of this project conducted by GenoMed4All and Synthema EU consortia is to link genomic information with transcriptomic data for possibly improving the prediction of clinical outcomes in MDS patients. Patients and Methods.Clinical, cytogenetic, genomic (somatic mutations screening of 31 target genes) and transcriptomic (bulk RNA-seq of CD34 + bone marrow cells) data were collected at diagnosis in 389 MDS patients. Transcriptomic and genomic profiles were processed and the former were normalized before Principal Component Analysis (PCA) dimensionality reduction to mine the interdependency of expression-wide perturbation and recurrent genomic alterations. The prognostic impacts of genetic, cytogenetic, transcriptomic, clinical and demographic features were assessed with a penalized Cox's proportional hazards model [Gerstung M et al, Nat Commun. 2015. 6, 5901] considering the Overall Survival (OS) as primary end point. A 5-fold cross-validating (CV) scheme was exploited to control bias in risk estimation. Model accuracy was assessed using Harrell's concordance index (C-index). An independent validation of the results on 202 patients was planned. Results.We first processed each data layer assessing data robustness, removed not informative variables and scaled quantitative ones. We considered recurrent genomic and cytogenetic lesions (present in ≥5 patients), platelets, hemoglobin and bone marrow blasts (%), age and sex as covariates. To explore the main patterns of expression changes, PCA was performed to reduce multidimensional correlated expression features (20 PCs was selected, explaining 42% of the total transcriptomic variability). To evaluate the prognostic power of each data layer we grouped all available features into five groups: gene mutations (n=15), cytogenetic alterations (n=7), expression data (n=20), blood counts (n=3) and demographic variables (n=2). Within a 5-fold CV we combined these variables in our integrative model to calculate MDS patients risk. The obtained predictive accuracy (C-index) for OS was 0.83, underlying that transcriptomic data significantly improved the current standard prognostic scoring systems. Accordingly, in our patient population, the C-index of the conventional IPSS-R score and the new IPSS-M were 0.68 and 0.76, respectively. A similar improvement by adding transcriptomic data was observed in prediction of the risk of AML evolution. Moreover, by analyzing the contribution of each feature category to the OS probability ( Figure 1), in term of explained variance, the relative impact of transcriptomic is 40%, with the remaining prognostic information distributed among genomic features (somatic gene mutations and cytogenetics lesions, 24%), demographics (20%) and clinical features (15%). An independent validation of these results on 202 patients is currently ongoing. Figure 2 shows an example of personalized survival prediction using patients from the study population. In two subjects with same clinical phenotype and mutations leading to a similar IPSS-M prognosis, the integrative model captures additional prognostic information and efficiently predicts clinical outcome. Given the complexity of our model, specific technological support is needed to combine data at individual patient level and to translate it into a personalized outcome prediction. To this aim, we created a prototype web portal based on our dataset for user-defined genomic/transcriptomic and clinical features. Conclusion. In predicting survival of MDS patients, genomic, transcriptomic and diagnostic clinical variables all have utility, with a significant contribution from the transcriptome.

Reclassification of Ascertain (ASTX727-02) Myelodysplastic Syndrome (MDS) Patients: Outcomes Including Clinical Response, Overall Survival (OS), and Leukemia Free Survival (LFS) Based on IPSS-R and IPSS-M Scoring Systems

Background: Oral decitabine/cedazuridine (ASTX727) is a fixed dose combination of decitabine (35 mg) and the cytidine deaminase inhibitor cedazuridine (100 mg) given once daily X 5 days on a 28-day cycle producing pharmacokinetic (PK) exposure equivalent to the standard intravenous (IV) decitabine regimen of 20 mg/m 2 daily X 5 days on a 28-day cycle. This was demonstrated in the pivotal ASCERTAIN study (Garcia-Manero, et al, ASH 2019) as it provided a PK bridge to existing decitabine data and demonstrated median OS was 31.7 mo. (Savona,et al. Intl. MDSF International Congress on MDS, 2021). Subjects were initially classified by the International Prognosis Scoring System (IPSS) for historical reasons, however IPSS-R and IPSS-M are enhancements to IPSS that provide dynamic risk assessment for predicting clinical outcomes in MDS. Here, the objective was to re-classify the MDS subjects enrolled on the ASCERTAIN study by IPSS-R and IPSS-M and measure the impact of informing patient outcomes based on re-calculated risk assessment. Methods: One hundred thirty-three subjects with MDS/CMML (chronic myelomonocytic leukemia) were enrolled in ASCERTAIN and were randomly assigned either IV decitabine for cycle 1 and oral decitabine/cedazuridine for cycle 2 or the opposite treatment sequence. All subjects continuing beyond cycle 2 received oral decitabine/cedazuridine for all subsequent cycles until treatment discontinuation for disease progression, toxicity, patient's decision, or HSCT. Whole blood collected prior to treatment was used for DNA isolation and molecular abnormalities identified using next-generation sequencing (NGS) hematologic malignancy panel of 179 genes including all genes commonly mutated in MDS. In the initial analysis of clinical outcomes, subjects were classified by IPSS with response assessment by IWG 2006 IPSS low and Int-1 risk levels were categorized as lower-risk MDS (LR-MDS), whereas IPSS Int-2 and high-risk categories were categorized as higher-risk MDS (HR-MDS). Subjects with sufficient data based on (e.g., available heme parameters, cytogenetics, NGS etc.) were reclassified by IPSS-R and IPSS-M. Subjects with IPSS-R score of ≤3.5 or IPSS-M of either very low, low, or moderate low were categorized as LR-MDS. Similarly subjects with IPSS-R score of >3.5 or an IPSS-M categorization of either moderate high, high, or very high were categorized as HR-MDS. Reclassified subjects were reassessed for CR (Complete Response), OS, and LFS and Harrell's concordance index (c-index) was used to describe the level of agreement between each scale and outcomes. Results: Based on the available data, the number of MDS subjects in the different risk classifications were the following: IPSS: 117, IPSS-R: 104, and IPSS-M: 105. Thirteen and 12 subjects on the IPSS could not be reclassified in the IPSS-R and IPSS-M, respectively, including 5 Int-1 and 2 low-risk MDS cases. CMML subjects were excluded from these analyses. Re-classification generally resulted in the upgrade of the subjects from LR-MDS to HR-MDS (Fig. 1). Thirty-one (26.5%) subjects from IPSS were reallocated in the IPSS-R to different risk categories; 3 (9.7%) were downgraded and 28 (90.3%) were upgraded. Similarly on reclassification with IPSS-M 34 (32.4%) of the patients reclassified; 5 (14.8%) were downgraded and 29 (85.3%) were upgraded. For IPSS LR-MDS, 23.2% of patients achieved CR, and 22.9% in the IPSS HR-MDS. Similarly, 21.6% patients achieved CR in the IPSS-R LR-MDS, and 23.9% in the IPSS-R HR-MDS; 26.3% patients achieved CR in the IPSS-M LR-MDS, and 20.9% in the IPSS-M HR-MDS. The c-index for the IPSS was .64 (OS) and .67 (LFS), IPSS-R was .70 (OS) and .71 (LFS), and .75 (OS) and .78 (LFS) for the IPSS-M (Table1). Conclusion: Reclassification from IPSS to IPSS-R or IPSS-M upgraded multiple subjects from a LR to a HR category, describing the ASCERTAIN patient population as a majority higher risk population with worse prognosis than previously assumed based on the IPSS. The efficacy as measured by the CR rates did not change when the LR and HR categories were defined by the different risk stratification systems. In contrast, the c-index improves with migration from IPSS to IPSS-R to IPSS-M, indicating an increased discriminatory ability of IPSS-M score in comparison to IPSS and IPSS-R, to predict patient outcomes.

A Prognostic Model Integrating Circulating Clonal Plasma Cells into R-ISS and PET-CT Based Tumor Burden in Transplant-Ineligible Multiple Myeloma

Introduction: Circulating clonal plasma cell (cCPC) is an easily obtained and non-invasive biomarker in multiple myeloma (MM), providing distinct prognostic information from the revised-international staging system (R-ISS). However, its prognostic relevance has been investigated primarily in transplant-eligible and clinical trial settings. Consequently, there is limited data on the impact of cCPC on outcomes and optimal cut-off values in transplant-ineligible (TIE) MM patients. Moreover, appropriate combinations to refine risk stratification, such as tumor burden measured by positron emission tomography-computed tomography (PET-CT), have yet to be fully explored. Methods: We performed a retrospective analysis on newly diagnosed TIE-MM patients with available peripheral blood (PB) flow cytometry (FCM) and PET-CT data from 2011 to 2021. Patients with primary plasma cell leukemia (n=8), defined as >5% monoclonal plasma cells in PB, were excluded from the study. Quantification of cCPC was performed using 6-color (2011-2016) or 8-color (2017-2021) FCM staining for CD19, CD56, CD38, CD138, cytoplasmic κ, and cytoplasmic λ to confirm monoclonality (Beckman Coulter, USA), which required at least 150000 mononuclear cell counts with the detectable limit at 0.01% (10 -4) based on previous reports (Gonsalves et al). Tumor burden was assessed by calculating total lesion glycolysis (TLG) using the following formula: the mean standardized uptake value (SUV) multiplied by the total tumor volume with an absolute SUV >2.5, using Metavol Software (Hokkaido University, Japan). Receiver operating curve analysis with the Youden index was employed to determine the optimal cutoff for predicting 2-year progression-free survival (PFS) and 5-year overall survival (OS). The predictive performance of R-ISS, cCPC, and TLG alone, along with their combinations, was compared using Harrell's concordance index (c-index). Results: 114 patients with TIE-MM were analyzed. The median age was 78 years (interquartile range [IQR] 73-83). The number of patients classified as R-ISS Ⅰ, Ⅱ, and Ⅲ was 12 (10.5%), 82 (71.9%), and 20 (17.5%), respectively. Most patients (85.1%) received proteasome inhibitor-based induction remission therapy. cCPCs were detectable in 90 patients (80.7%), with a median level of 6 ×10 -4 (IQR 1-22). Regarding tumor burden, 73 patients (64%) had at least one fluorodeoxyglucose avid lesion with a median TLG of 34.7 (IQR 0-268). cCPC levels were significantly correlated with CPC in bone marrow (R =0.24, P =0.01) but not with TLG (R =-0.01, P =0.88) and involved free light chain levels (R =0.07, P =0.41). The optimal cut-off values in predicting 2-year PFS and 5-year OS were 9 and 10 ×10 -4 for cCPC, and 325 and 343 for TLG, respectively. For simplicity, we classified >10 ×10 -4 for cCPC and >300 for TLG as high-risks. Notably, approximately two-thirds of the patients (62.3%) had at least one high-risk feature (Figure 1). Both high cCPC and TLG showed significant associations with worse OS and PFS (5-year OS rate: cCPC high vs. low, 40.2% vs. 69.3%; TLG high vs. low, 38% vs. 64.5%; 2-year PFS rate: cCPC high vs. low, 45.7% vs. 69.1%; TLG high vs. low, 46.4% vs. 64% all P < 0.05). These factors retained prognostic significance after adjustment for R-ISS on multivariate analysis. Finally, we constructed a novel predictive model integrating RISS, cCPC, and TLG. This model discriminated OS among patients with 0 (5-year OS rate, 80% [95% confidence interval 68.4%-93.6%]), 1 (50.5% [37.7%-67.7%]), and ≧2 factors (29% [13.3%-63.5%], P <0.001, Figure 2). The c-index was the highest (0.651) in this combined model compared to the single use of these risk factors (0.585, 0.594, and 0.569). Conclusions: We demonstrated the prognostic impact of cCPC, which is independently associated with R-ISS and TLG, on real-world TIE-MM patients. The new prediction model, integrated with these different clinical and biological features, may provide robust outcome discrimination.

Nomogram based on pretreatment hepatic and renal function indicators for survival prediction of locally advanced esophageal squamous cell carcinoma with treatment of neoadjuvant chemoradiotherapy plus surgery.

The parameters for survival prediction of esophageal squamous cell carcinoma (ESCC) patients treated with neoadjuvant chemoradiotherapy (NCRT) combined with surgery are unclear. Here, we aimed to construct a nomogram for survival prediction of ESCC patients treated with NCRT combined with surgery based on pretreatment serological hepatic and renal function tests. A total of 174 patients diagnosed as ESCC were enrolled as a training cohort from July 2007 to June 2019, and approximately 50% of the cases (n = 88) were randomly selected as an internal validation cohort. Univariate and multivariate Cox survival analyses were performed to identify independent prognostic factors to establish a nomogram. Predictive accuracy of the nomogram was evaluated by Harrell's concordance index (C-index) and calibration curve. ALT, ALP, TBA, TP, AST, TBIL and CREA were identified as independent prognostic factors and incorporated into the construction of the hepatic and renal function test nomogram (HRFTNomogram). The C-index of the HRFTNomogram for overall survival (OS) was 0.764 (95% CI 0.701-0.827) in the training cohort, which was higher than that of the TNM staging system (0.507 (95% CI 0.429-0.585), P < 0.001). The 5-year OS calibration curve of the training cohort demonstrated that the predictive accuracy of the HRFTNomogram was satisfactory. Moreover, patients in the high-risk group stratified by the HRFTNomogram had poorer 5-year OS than those in the low-risk group in the training cohort (27.4% vs. 80.3%, P < 0.001). Similar results were observed in the internal validation cohort. A novel HRFTNomogram might help predict the survival of locally advanced ESCC patients treated with NCRT followed by esophagectomy.

Validation of non-muscle-invasive bladder cancer risk stratification updated in the 2021 European Association of Urology guidelines.

The objective of this study is to validate the predictive ability of the 2021 European Association of Urology (EAU) risk model compared to that of existing risk models, including the 2019 EAU model and risk scoring tables of the European Organization for Research and Treatment of Cancer, Club Urologico Espanol de Tratamiento Oncologico,and Japanese Nishinihon Uro-oncology Extensive Collaboration Group. This retrospective multi-institutional database study included two cohorts-3024 patients receiving intravesical bacillus Calmette-Guerin (BCG) treatment (BCG cohort) and 789 patients not receiving BCG treatment (non-BCG cohort). The Kaplan-Meier estimate and log-rank test were used to visualize and compare oncological survival outcomes after transurethral surgery among the risk groups. Harrell's concordance index (C-index) was used to evaluate the predictive ability of the models. We observed a risk shift from the 2019 EAU risk grouping to the 2021 EAU risk grouping in a substantial number of patients. For progression, the C-index of the 2021 EAU model was significantly higher than that of the 2019 EAU model in both the BCG (0.617 vs. 0.572; P = 0.011) and non-BCG (0.718 vs. 0.560; P < 0.001) cohorts. According to the 2021 EAU model, 731 (24%) and 130 (16%) patients in the BCG and non-BCG cohorts, respectively, were considered to have a very high risk. Survival analysis showed no significant differences among the five very high-risk subgroups in both cohorts. A major limitation was potential selection bias owing to the retrospective nature of this study. The updated 2021 EAU model showed better stratification than the three existing risk models, especially for progression, in both cohorts, determining the most appropriate postoperative treatment and identifying patients requiring intensified surveillance or early cystectomy.

Machine Learning Approach to Understand Real-World Treatment in Patients with Higher-Risk Myelodysplastic Syndromes

Background: Higher-risk myelodysplastic syndromes (HR-MDS) is a severe and aggressive form of MDS that represents a group of rare and heterogeneous hematologic disorders that primarily affect the bone marrow and blood cells. Treatment typically involves a combination of supportive care and active treatment with targeted therapies, hypomethylating agents, chemotherapy, and hematopoietic stem cell transplant. In this study, we have evaluated real-world treatment patterns in patients with HR-MDS with machine-learning (ML) models to identify clinical and demographic features that may impact decisions about which patients receive systemic therapies and when the therapies are implemented. Methods: We identified 821 patients aged ≥18 years with HR-MDS diagnosed between January 2015 and April 2022, using ConcertAI's RWD360 database linked with open claims data. RWD360 consists of structured records from US-based oncology electronic health record systems. Demographic and clinical data up to 1 year prior to the date of HR-MDS diagnosis were used to develop a family of XGBoost-based ML models to investigate drivers of treatment. We created five models: a time-to-therapy model that predicts patient-level hazard ratios and four classification models that predict likelihood of therapy in the first 2, 3, 6, and 12 months, respectively, following HR-MDS diagnosis (therapy yes/no models). Model performance was evaluated using Harrell's concordance index (C-index) and Akaike information criterion (AIC) for the time-to-therapy model and area under curve (AUC) for the therapy yes/no models. Key features included laboratory values, Charlson comorbidity index (CCI), and prescribed medications. To better capture the temporal characteristics of a patient's disease state evolution, clinical features were organized into time windows alongside the variations within them. We used recursive feature elimination to bring the total features used in the model to 50 to enhance the interpretability and compactness. Five-fold cross-validation (data was split 5 ways; for each fold, a different 20% is left out for validation, while 80% of the data was used to train the model) on the full 821-patient cohort was performed to train and evaluate the model. Results: The median age in this patient cohort was 73 years (range = 21-88 years); approximately 60% were male, and 66% patients were treated in community oncology practices. Only 42.5% patients were recorded to ever receive systemic active therapy, and of those, the median time to therapy was 4.2 months (Figure 1). Patients had a median hemoglobin concentration of 9.4 g/dL (range = 4.4-18.9 g/dL), median CCI of 1 (range = 0-10), and median polypharmacy of 10 (computed on patients who had at least 1 concomitant medication (49%); range = 1-93) around time of diagnosis (assessed 1 year prior, and up to 30 days after diagnosis date). Model performance for the time-to-therapy model (C-index = 0.78; AIC = 3314.2) and for the therapy yes/no models (AUC = 0.7-0.74) were evaluated. A key predictive feature across models was the minimum laboratory value of white blood cell (WBC) count near the point of diagnosis; low WBC values around time of diagnosis was consistently a strong predictor for receiving systemic therapy and receiving it earlier. Two other top predictors of systemic therapy and its timing were the rate of fall in measurements of blood glucose and the rate of fall of hemoglobin during the 180 to 30 days before diagnosis. Table 1 details model performance and the top 3 features impacting time and likelihood of therapy for all 5 models. Conclusion: We have built a family of data-driven ML models for patients with HR-MDS, based on data collected in routine clinical practice, towards predicting time-to and probability-of first-line systemic therapy with reasonable accuracy driven by patient characteristics. A majority of patients diagnosed with HR-MDS in this study appeared to receive delayed or no systemic therapy in their treatment journey. To bridge gaps in care and facilitate timely treatment in eligible patients, data-driven predictive modeling that take into account detailed patient-level factors is critical. Further analysis on outcomes of untreated patients linked with their key clinical characteristics identified using these models may provide opportunities to improve clinical care by implementing strategies for treatments based on patient-level factors.

Lower Prognostic Performance of the International Metabolic Prognostic Index Compared to the Conventional IPI and NCCN-IPI in a Real-World Cohort of Diffuse Large B-Cell Lymphoma

Introduction: The international metabolic prognostic index (IMPI) is a recently proposed model that shows improved outcome prediction for patients with diffuse large B-cell lymphoma (DLBCL) compared to the International Prognostic Index (IPI) (Mikhaeel et al.). IMPI combines total metabolic tumor volume (TMTV) as continuous variables with clinical factors, allowing individualized prognosis estimation. However, the applicability of IMPI to routine clinical practice remains uncertain since the model construction was based on data from clinical trials. Furthermore, the comparison of the model performance between IMPI and the National Comprehensive Cancer Network IPI (NCCN-IPI) is not reported so far. Methods: We performed retrospective analysis of consecutive patients with newly diagnosed DLBCL who were initially treated with R-CHOP (rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone) at Kameda Medical Center between 2006 and 2020. Patients who had no fluorodeoxyglucose-avid lesions were excluded. TMTV was defined as the volume of lymphoma-associated lesions with a standardized uptake value of ≥4 as the absolute threshold (Metavol, Hokkaido University, Japan). According to the original study (Mikhaeel et al.), IMPI was calculated using the following formula: 0.003077330 × (MTV lower than the median of 307.9ml) - 0.002761985 × (MTV higher than the median of 307.9ml) + 0.008092449 × Age -0.114645415 × Stage 2 + 0.281141117 × Stage 3 + 0.322247142 × Stage 4. The receiver operating curve (ROC) for predicting 5-year survival was compared among IMPI-, IPI-, and NCCN-IPI-based classifications. The predictive performance of these three systems was evaluated using the Akaike information criterion (AIC) and Harrell's concordance index (c-index). Results: A total of 314 patients were included in the study. The median age was 72 years old (interquartile range [IQR] 64-79), which was younger than the original data (median 62, IQR 51-70). Regarding the distributions of IPI risk groups, this cohort had a lower number of low-risk patients (27.0% vs. 32.4%) and a higher number of high-risk patients (23.9% vs. 19.5%). The median TMTV was 168 (IQR 47-530), which was approximately two times lower than that in the original study (median 307, IQR 77-838). After calculating and ranking the individual IMPI scores, nearly half of the patients (n=147, 46.8%) were reclassified into different risk groups from the original IPI categories (Figure 1); 76 (51.7%) were upstaged, and 71 (48.3%) were downstaged. Of note, the reclassification resulted in a significantly better OS prediction only for the IPI high-risk group (5-year OS rate, IMPI high-risk vs. IMPI not high-risk, 41.9% vs. 59%, P =0.042), while it did not for the IPI low-, low-intermediate-, and high-intermediate-risk groups. The area under the curve (AUC) from the ROC curve predicting 5-year survival was not significantly different between IMPI- (AUC 0.65) and either IPI- (AUC 0.67, P =0.42) or NCCN-IPI-categories (AUC 0.68, P=0.25). Finally, the predictive performance and model fitness were compared among these three indices. IMPI had the highest AIC (1117.68) and the lowest c-index (0.643) for OS prediction, whereas NCCNIPI yielded the best performance (AIC 1097.66 and c-index 0.677) (Table 1). Conclusions: We demonstrated that the IMPI model did not outperform the conventional IPI and NCCN-IPI systems in a real-world cohort of DLBCL patients. The difference in age and tumor burden, which are the major components of IMPI, may lead to poorer prognostic performance. Further large-scale studies are warranted to validate the prognostic value of IMPI outside of clinical trials.

Plaque enhancement predicts recurrence in acute ischemic stroke patients with large artery intracranial atherosclerosis

BackgroundThe association between the degree of plaque enhancement and ischemic brain stroke recurrence remains unclear. We aimed to establish models to predict plaque enhancement and stroke recurrence. MethodsSeventy-eight participants with acute ischemic brain stroke due to intracranial arterial stenosis were recruited and divided into high enhancement (HE) and non-HE groups. The relationship between imaging characteristics (degree of stenosis, minimal lumen area, intraplaque hemorrhage, and plaque burden) and the degree of plaque contrast enhancement was analyzed. Inflammatory cytokine expression was examined by flow cytometry. Independent predictors of stroke recurrence were investigated via multivariate Cox proportional hazards regression analysis. Nomogram was used to construct a prediction model. Harrell's concordance indices (c-indices) and calibration curves were used to assess the discrimination of the nomogram. A risk prediction nomogram for prognosis was constructed. ResultsThirty-three participants were assigned to the HE group and 45 to the non-HE group. The degree of stenosis and plaque burden in the HE group was higher than that in the non-HE group (P<0.05). Multiple linear regression analysis showed the degree of stenosis was associated with HE (β=0.513; P=0.000). After adjusting for confounding factors, age (HR=1.115; 95%CI=1.034–1.203, P=0.005) and HE plaques (HR=10.457; 95%CI=1.176–93.018; P=0.035) were independent risk factors of stroke recurrence, whereas cytokine levels were not statistically significant between two group. ConclusionsHE of intracranial atherosclerosis plaques is an independent factor for ischemic brain stroke recurrence.

A simplified frailty index and nomogram to predict the postoperative complications and survival in older patients with upper urinary tract urothelial carcinoma.

This study was designed to investigate the clinical value of a simplified five-item frailty index (sFI) for predicting short- and long-term outcomes in older patients with upper urinary tract urothelial carcinoma (UTUC) patients after radical nephroureterectomy (RNU). This retrospective study included 333 patients (aged ≥65 years) with UTUC. Patients were classified into five groups: 0, 1, 2, 3, and 3+, according to sFI score. The variable importance and minimum depth methods were used to screen for significant variables, and univariable and multivariable logistic regression models applied to investigated the relationships between significant variables and postoperative complications. Survival differences between groups were analyzed using Kaplan-Meier plots and log-rank tests. Cox proportional hazards regression was used to evaluate risk factors associated with overall survival (OS) and cancer-specific survival (CSS). Further, we developed a nomogram based on clinicopathological features and the sFI. The area under the curve (AUC), Harrel's concordance index (C-index), calibration curve, and decision curve analysis (DCA) were used to evaluate the nomogram. Of 333 cases identified, 31.2% experienced a Clavien-Dindo grade of 2 or greater complication. Random forest-logistic regression modeling showed that sFI significantly influenced the incidence of postoperative complications in older patients (AUC= 0.756). Compared with patients with low sFI score, those with high sFI scores had significantly lower OS and CSS (p < 0.001). Across all patients, the random survival forest-Cox regression model revealed that sFI score was an independent prognostic factor for OS and CSS, with AUC values of 0.815 and 0.823 for predicting 3-year OS and CSS, respectively. The nomogram developed was clinically valuable and had good ability to discriminate abilities for high-risk patients. Further, we developed a survival risk classification system that divided all patients into high-, moderate-, and low-risk groups based on total nomogram points for each patient. A simple five-item frailty index may be considered a prognostic factor for the prognosis and postoperative complications of UTUC following RNU. By using this predictive model, clinicians may increase their accuracy in predicting complications and prognosis and improve preoperative decision-making.

The T Cell Immunoscore as a Reference for Biomarker Development Utilizing Real-World Data from Patients with Advanced Malignancies Treated with Immune Checkpoint Inhibitors.

We aimed to determine the prognostic value of an immunoscore reflecting CD3+ and CD8+ T cell density estimated from real-world transcriptomic data of a patient cohort with advanced malignancies treated with immune checkpoint inhibitors (ICIs) in an effort to validate a reference for future machine learning-based biomarker development. Transcriptomic data was collected under the Total Cancer Care Protocol (NCT03977402) Avatar® project. The real-world immunoscore for each patient was calculated based on the estimated densities of tumor CD3+ and CD8+ T cells utilizing CIBERSORTx and the LM22 gene signature matrix. Then, the immunoscore association with overall survival (OS) was estimated using Cox regression and analyzed using Kaplan-Meier curves. The OS predictions were assessed using Harrell's concordance index (C-index). The Youden index was used to identify the optimal cut-off point. Statistical significance was assessed using the log-rank test. Our study encompassed 522 patients with four cancer types. The median duration to death was 10.5 months for the 275 participants who encountered an event. For the entire cohort, the results demonstrated that transcriptomics-based immunoscore could significantly predict patients at risk of death (p-value < 0.001). Notably, patients with an intermediate-high immunoscore achieved better OS than those with a low immunoscore. In subgroup analysis, the prediction of OS was significant for melanoma and head and neck cancer patients but did not reach significance in the non-small cell lung cancer or renal cell carcinoma cohorts. Calculating CD3+ and CD8+ T cell immunoscore using real-world transcriptomic data represents a promising signature for estimating OS with ICIs and can be used as a reference for future machine learning-based biomarker development.

Development of a Survival Nomogram for Esophageal Squamous Cell Carcinoma Patients: a Population-Based Analysis.

In this study, we developed a prognostic nomogram for esophageal squamous cell carcinoma (ESCC) patients. Patients diagnosed with ESCC from the Surveillance, Epidemiology, and End Results (SEER) database (1975-2017) and a local hospital were enrolled in this retrospective cohort study. Prognoses were analyzed using the R language software, and the predictive power of the model was then assessed by the Harrell concordance index (C-index) and the area under the curve (AUC) of the receiver operating characteristic (ROC) curve. In total, 2915 ESCC patients from SEER database were divided into training and validation cohorts. Multivariate analysis revealed that sex, marital status, tumor-node-metastasis (TNM) stage, surgery, chemotherapy, and radiation all showed a significant association with overall survival (OS) and cancer-specific survival (CSS) (also with tumor grade). These characteristics were employed to build a nomogram. The C-index of the nomogram for OS and CSS prediction was 0.743 and 0.748 for the training cohort, which were superior to the predictive power of the 7th TNM staging system. The AUCs of the nomogram for predicting 2- and 5-year OS were 0.805 and 0.812, respectively, and the AUCs for CSS were 0.811 and 0.821, respectively. ROC and calibration curves of data from the SEER internal validation set and of data from our hospital showed that this model had good accuracy for predicting the prognosis of ESCC patient. The nomogram developed in this study provides a useful tool for accurately estimating OS and CSS for ESCC patients.

A cellular senescence-related genes model allows for prognosis and treatment stratification of cervical cancer: a bioinformatics analysis and external verification.

Cervical cancer (CC) is highly lethal and aggressive with an increasing trend of mortality for females. Molecular characterization-based methods hold great promise for improving the diagnostic accuracy and for predicting treatment response. The mRNAs expression data of CC patients and cellular senescence-related genes were obtained from the Cancer Genome Atlas (TCGA) and CellAge databases, respectively. Differentially expressed genes (DEGs) of senescence related genes between tumor and normal tissues were used for Least absolute shrinkage and selection operator (LASSO) regression to construct a prognostic model. Univariate and LASSO regression analyses were applied to establish a predictive nomogram. The performance of the nomogram were evaluated by Kaplan-Meier curve, receiver operating characteristic (ROC), Harrell's concordance index (C-index), and calibration curve. GSE44001 and GSE52903 were used for external validation. We established a cellular senescence-related genes-based stratified model, and a multivariable-based nomogram, which could accurately predict the prognosis of CC patients in the TCGA database. The Kaplan-Meier curve indicated that patients in the low-risk group had considerably better overall survival (OS, P =2.021e-05). The area under the ROC curve (AUC) of this model was 0.743 for OS. Multivariate analysis found that the 6-gene risk signature (HR=3.166, 95%CI: 1.660-6.041, P<0.001) was an independent risk factor for CC patients. We then designed an OS-associated nomogram that included the risk signature and clinicopathological factors. The AUC reached 0.860 for predicting 5-year OS. The nomogram showed excellent consistency between the predictions and actual survival observations. Two external GEO validations were corresponding to the gene expression pattern in TCGA. Our results suggested a six-senescence related signature and established a prognostic nomogram that reliably predicted the overall survival for CC. These findings may be beneficial to personalized treatment and medical decision-making.

Utility of the revised FIGO2023 staging with molecular classification in endometrial cancer

ObjectivesMolecular classification was introduced in endometrial cancer staging following the transition of the International Federation of Gynecology and Obstetrics (FIGO) 2008 to FIGO2023. In the early stages, p53 abnormal endometrial carcinoma with myometrial involvement was upstaged to stage IICm, in addition to the downstaging of POLE mutation endometrial cancer to stage IAm. This study compared the goodness of fit and discriminatory ability of FIGO2008, FIGO2023 without molecular classification (FIGO2023), and FIGO2023 with molecular classification (FIGO2023m); no study has been externally validated to date. MethodsThe study included 265 patients who underwent initial surgery at the National Cancer Center Hospital between 1997 and 2019 and were pathologically diagnosed with endometrial cancer. The three classification systems were compared using Harrell's concordance index (C-index), Akaike information criterion (AIC), and time-dependent receiver operating characteristic (ROC) curves. A higher C-index score and a lower AIC value indicated a more accurate model. ResultsAmong the three classification systems, FIGO2023m had the lowest AIC value (FIGO2023m: 455.925; FIGO2023: 459.162; FIGO2008: 457.901), highest C-index (FIGO2023m: 0.768; FIGO2023: 0.743; FIGO2008: 0.740), and superior time-dependent ROC curves within 1 year after surgical resection. In the stage IIIC, patients with p53 abnormalities had considerably lower 5-year overall survival than those with a p53 wild-type pattern (24.3% vs. 83.7%, p = 0.0005). ConclusionsFIGO2023m had the best discriminatory ability compared with FIGO2008 and FIGO2023. Even in advanced stages, p53 status was a poor prognostic factor. When feasible, molecular subtypes can be added to the staging criteria to allow better prognostic prediction in all stages.

Development and Validation of an XGBoost-Algorithm-Powered Survival Model for Predicting In-Hospital Mortality Based on 545,388 Isolated Severe Traumatic Brain Injury Patients from the TQIP Database.

Traumatic brain injury (TBI) represents a significant global health issue; the traditional tools such as the Glasgow Coma Scale (GCS) and Abbreviated Injury Scale (AIS) which have been used for injury severity grading, struggle to capture outcomes after TBI. This paper aims to implement extreme gradient boosting (XGBoost), a powerful machine learning algorithm that combines the predictions of multiple weak models to create a strong predictive model with high accuracy and efficiency, in order to develop and validate a predictive model for in-hospital mortality in patients with isolated severe traumatic brain injury and to identify the most influential predictors. In total, 545,388 patients from the 2013-2021 American College of Surgeons Trauma Quality Improvement Program (TQIP) database were included in the current study, with 80% of the patients used for model training and 20% of the patients for the final model test. The primary outcome of the study was in-hospital mortality. Predictors were patients' demographics, admission status, as well as comorbidities, and clinical characteristics. Penalized Cox regression models were used to investigate the associations between the survival outcomes and the predictors and select the best predictors. An extreme gradient boosting (XGBoost)-powered Cox regression model was then used to predict the survival outcome. The performance of the models was evaluated using the Harrell's concordance index (C-index). The time-dependent area under the receiver operating characteristic curve (AUC) was used to evaluate the dynamic cumulative performance of the models. The importance of the predictors in the final prediction model was evaluated using the Shapley additive explanations (SHAP) value. On average, the final XGBoost-powered Cox regression model performed at an acceptable level for patients with a length of stay up to 250 days (mean time-dependent AUC = 0.713) in the test dataset. However, for patients with a length of stay between 20 and 213 days, the performance of the model was relatively poor (time-dependent AUC < 0.7). When limited to patients with a length of stay ≤20 days, which accounts for 95.4% of all the patients, the model achieved an excellent performance (mean time-dependent AUC = 0.813). When further limited to patients with a length of stay ≤5 days, which accounts for two-thirds of all the patients, the model achieved an outstanding performance (mean time-dependent AUC = 0.917). The XGBoost-powered Cox regression model can achieve an outstanding predictive ability for in-hospital mortality during the first 5 days, primarily based on the severity of the injury, the GCS on admission, and the patient's age. These variables continue to demonstrate an excellent predictive ability up to 20 days after admission, a period of care that accounts for over 95% of severe TBI patients. Past 20 days of care, other factors appear to be the primary drivers of in-hospital mortality, indicating a potential window of opportunity for improving outcomes.

Nomogram for predicting overall survival in patients with large (>5 cm) hepatocellular carcinoma based on real-world practice.

Patients with large (>5 cm) hepatocellular carcinoma (HCC) have limited treatment options, thus necessitating the identification of prognostic factors and the development of predictive tools. This study aimed to identify prognostic factors and to construct a nomogram to predict survival outcomes in patients with large HCC. A cohort of 438 patients, who were diagnosed with large HCC at a tertiary hospital between 2015 and 2018, was analyzed. Cox proportional hazards models were used to identify key prognosticators of overall survival (OS), and an independent set of prognostic factors was used to develop a nomogram. The discrimination and calibration abilities of the nomogram were assessed and internal validation was performed using cross-validation and bootstrapping methods. During a median follow-up of 9.3 months, the median OS was 9.9 months, and the 1-year OS rate was 43.9%. Multivariable Cox regression analysis revealed that performance status, modified albumin-bilirubin grade, tumor size, extent of portal vein tumor thrombosis, and initial treatment significantly affected OS. The newly developed nomogram incorporating these variables demonstrated favorable accuracy (Harrell's concordance index, 0.807). The newly developed nomogram facilitated the estimation of individual survival outcomes in patients with large HCC, providing an acceptable level of accuracy.

Improving the assessment of malnutrition in cancer: Using systemic inflammation markers as a supplement to the inflammation items of the GLIM criteria

Systemic inflammation is a key pathogenic criterion for diagnosing malnutrition using the Global Leadership Initiative on Malnutrition (GLIM) criteria. Although cancer is commonly considered as a chronic inflammation-related disease, the inflammatory burden may vary depending on the type and stage of cancer. Therefore, a more precise definition of inflammation criteria could facilitate the identification of malnutrition in cancer. This prospective multicenter study included 1683 cancer patients screened via NRS2002 for malnutrition risk. The inflammatory burden index (IBI), C-reactive protein (CRP) level, neutrophil-to-lymphocyte ratio (NLR), and albumin (ALB) level were used to assess the inflammatory burden. Kaplan-Meier and Cox regression analyses were used to determine the relationship between the GLIM criteria and overall survival. Harrell's concordance index (C-index) was used to compare the discriminative performance of the original, IBI-based, CRP-based, NLR-based, and ALB-based GLIM criteria for survival. Logistic regression models were used to assess the association between GLIM criteria and short-term outcomes, length of hospital stay, and hospitalization costs. Compared to the original GLIM criteria, the IBI/CRP/NLR/ALB-based GLIM criteria better predicted the long-term outcomes of patients with cancer (chi-square: 1.316 vs. 78.321 vs. 74.740 vs. 88.719 vs. 100.921). The C-index revealed that the inflammation marker-based GLIM criteria showed significantly better prognostic accuracy than the original GLIM criteria. The ALB-based GLIM criteria exhibited the best prognostic accuracy. The inflammation marker-based GLIM criteria were independent predictive factors for the long-term prognosis of cancer. Patients with malnutrition had a 45% higher risk of adverse long-term prognoses than those without malnutrition. The inflammation marker-based GLIM criteria had good prognostic ability to predict outcomes at 3, 6, and 12 months. The stepwise effect of the grading of severity via the IBI-based GLIM criteria and CRP-based GLIM criteria was notable. The inflammation marker-based GLIM criteria are useful for predicting short-term outcomes, length of hospitalization, and hospitalization costs. The inflammation marker-based GLIM criteria have a stronger predictive value than the original GLIM criteria in evaluating both the short- and long-term prognoses of cancer patients. It is recommended to use the inflammation marker-based GLIM criteria for nutritional evaluation of cancer patients.

A novel postoperative nomogram and risk classification system for individualized estimation of survival among patients with parotid gland carcinoma after surgery.

Parotid gland carcinoma (PGC) is a rare but aggressive head and neck cancer, and the prognostic model associated with survival after surgical resection has not yet been established. This study aimed to construct a novel postoperative nomogram and risk classification system for the individualized prediction of overall survival (OS) among patients with resected PGC. Patients with PGC who underwent surgery between 2004 and 2015 from the Surveillance, Epidemiology, and End Results (SEER) database were randomized into training and validation cohorts (7:3). A nomogram developed using independent prognostic factors based on the results of the multivariate Cox regression analysis. Harrell's concordance index (C-index), time-dependent area under the curve (AUC), and calibration plots were used to validate the performance of the nomogram. Moreover, decision curve analysis (DCA) was performed to compare the clinical use of the nomogram with that of traditional TNM staging. In this study, 5077 patients who underwent surgery for PGC were included. Age, sex, marital status, tumor grade, histology, TNM stage, surgery type, radiotherapy, and chemotherapy were independent prognostic factors. Based on these independent factors, a postoperative nomogram was developed. The C-index of the proposed nomogram was 0.807 (95% confidence interval 0.797-0.817). Meanwhile, the time-dependent AUC (> 0.8) indicated that the nomogram had a satisfactory discriminative ability. The calibration curves showed good concordance between the predicted and actual probabilities of OS, and DCA curves indicated that the nomogram had a better clinical application value than the traditional TNM staging. Moreover, a risk classification system was built that could perfectly classify patients with PGC into three risk groups. This study constructed a novel postoperative nomogram and corresponding risk classification system to predict the OS of patients with PGC after surgery. These tools can be used to stratify patients with high or low risk of mortality and provide high-risk patients with more directed therapies and closer follow-up.

Establishment and validation of a novel nomogram to predict overall survival after radical nephrectomy

Objective: To establish a nomogram prognostic model for predicting the 5-, 10-, and 15-year overall survival (OS) of non-metastatic renal cell carcinoma patients managed with radical nephrectomy (RN), compare the modelled results with the results of pure pathologic staging, the Karakiewicz nomogram and the Mayo Clinic Stage, Size, Grade, and Necrosis (SSIGN) score commonly used in foreign countries, and stratify the patients into different prognostic risk subgroups. Methods: A total of 1 246 non-metastatic renal cell carcinoma patients managed with RN in Sun Yat-sen University Cancer Center (SYSUCC) from 1999 to 2020 were retrospectively analyzed. Multivariate Cox regression analysis was used to screen the variables that influence the prognosis for nomogram establishment, and the bootstrap random sampling was used for internal validation. The time-receiver operating characteristic curve (ROC), the calibration curve and the clinical decision curve analysis (DCA) were applied to evaluate the nomogram. The prediction efficacy of the nomogram and that of the pure pathologic staging, the Karakiewicz nomogram and the SSIGN score was compared through the area under the curve (AUC). Finally, patients were stratified into different risk subgroups according to our nomogram scores. Results: A total of 1 246 patients managed with RN were enrolled in this study. Multivariate Cox regression analysis showed that age, smoking history, pathological nuclear grade, sarcomatoid differentiation, tumor necrosis and pathological T and N stages were independent prognostic factors for RN patients (all P<0.05). A nomogram model named SYSUCC based on these factors was built to predict the 5-, 10-, and 15-year survival rate of the participating patients. In the bootstrap random sampling with 1 000 iterations, all these factors occurred for more than 800 times as independent predictors. The Harrell's concordance index (C-index) of SYSUCC was higher compared with pure pathological staging [0.770 (95% CI: 0.716-0.823) vs 0.674 (95% CI: 0.621-0.728)]. The calibration curve showed that the survival rate as predicted by the SYSUCC model simulated the actual rate, while the clinical DCA showed that the SYSUCC nomogram has a benefit in certain probability ranges. In the ROC analysis that included 857 patients with detailed pathological nuclear stages, the nomogram had a larger AUC (5-/10-year AUC: 0.823/0.804) and better discriminating ability than pure pathological staging (5-/10-year AUC: 0.701/0.658), Karakiewicz nomogram (5-/10-year AUC: 0.772/0.734) and SSIGN score (5-/10-year AUC: 0.792/0.750) in predicting the 5-/10-year OS of RN patients (all P<0.05). In addition, the AUC of the SYSUCC nomogram for predicting the 15-year OS (0.820) was larger than that of the SSIGN score (0.709), and there was no statistical difference (P<0.05) between the SYSUCC nomogram, pure pathological staging (0.773) and the Karakiewicz nomogram (0.826). The calibration curve was close to the standard curve, which indicated that the model has good predictive performance. Finally, patients were stratified into low-, intermediate-, and high-risk subgroups (738, 379 and 129, respectively) according to the SYSUCC nomogram scores, among whom patients in intermediate- and high-risk subgroups had a worse OS than patients in the low-risk subgroup (intermediate-risk group vs. low-risk group: HR=4.33, 95% CI: 3.22-5.81, P<0.001; high-risk group vs low-risk group: HR=11.95, 95% CI: 8.29-17.24, P<0.001), and the high-risk subgroup had a worse OS than the intermediate-risk group (HR=2.63, 95% CI: 1.88-3.68, P<0.001). Conclusions: Age, smoking history, pathological nuclear grade, sarcomatoid differentiation, tumor necrosis and pathological stage were independent prognostic factors for non-metastasis renal cell carcinoma patients after RN. The SYSUCC nomogram based on these independent prognostic factors can better predict the 5-, 10-, and 15-year OS than pure pathological staging, the Karakiewicz nomogram and the SSIGN score of patients after RN. In addition, the SYSUCC nomogram has good discrimination, agreement, risk stratification and clinical application potential.