An internally validated new clinical and inflammation-based prognostic score for patients with advanced hepatocellular carcinoma treated with sorafenib

Clinicopathological Aspects and Inflammation-Immune Markers in Alcohol and/or Hepatitis C Virus-Induced Hepatocellular Carcinoma Patients Treated With Sorafenib

Background

Tyrosine kinase inhibitors have been used to treat hepatocellular carcinoma (HCC), but the outcomes of patients under treatment vary. Since the roles of clinicopathological aspects and markers of chronic inflammation/immune homeostasis in the outcome of HCC patients treated with sorafenib are still unclear, these were the aims of this study.

Methods

Patients with alcohol-induced and/or hepatitis C virus (HCV)-induced HCC (n = 182) uniformly treated with sorafenib were included in the study. Baseline clinicopathological aspects of patients were computed from the medical records. The neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocyte ratio (LMR), systemic inflammation response index (SIRI), and systemic immune-inflammation index (SII) were obtained from the hematological exam performed before the administration of sorafenib. Overall survival (OS) was analyzed using Kaplan-Meier probabilities, log-rank test, and univariate and multivariate Cox proportional hazard ratio (HR) analyses.

Results

In multivariate analysis, alpha-foetoprotein (AFP) level and Child-Pugh score were predictors of OS. Patients with AFP levels higher than 157 ng/mL and Child-Pugh B or C had 1.40 (95% confidence interval (CI): 1.03 - 1.91, P = 0.03) and 1.64 (95% CI: 1.07 - 2.52, P = 0.02) more chances of evolving to death than the remaining patients, respectively. NLR, PLR, LMR, SIRI, and SII did not alter the OS of HCC patients.

Conclusions

AFP level and Child-Pugh score act as independent prognostic factors in patients with alcohol and/or HCV-induced HCC treated with sorafenib, but markers of chronic inflammation/immune homeostasis seem not to alter the outcome of patients with HCC induced by alcohol and/or HCV.

Development of a Multivariate Prognostic Model for Lenvatinib Treatment in Hepatocellular Carcinoma

BACKGROUND

Lenvatinib is a first-line agent for advanced hepatocellular carcinoma (HCC), but individual responses to treatment are highly heterogeneous. The aim of this study was to investigate the clinical parameters that influence the efficacy of Lenvatinib and to develop a prognostic model.

METHODS

We retrospectively enrolled 333 Lenvatinib-treated patients with HCC with a median age of 57 years. Two hundred nd sixty-three of these patients had BCLC (2022) stage C. The median overall survival (mOS) time within the cohort was 12.1 months, and the median progression-free survival (mPFS) time was 4.7 months. Univariate Cox regression, best subset regression, and Lasso regression were used to screen primary variables for possible contribution to OS, multivariate Cox analysis was used to fit selected models, and the final model was selected using the maximum area under the curve (AUC) and minimum AIC. Receiver operating curves (ROC), calibration curves, and decision curve analysis were plotted to assess model performance, and 5-fold cross-validation was performed for internal validation. X-tile software was used to select the best cutoff points and to divide the study cohort into 3 different risk groups.

RESULTS

Seven variables were included in the final model: BCLC stage, prior transarterial chemoembolization and immunotherapy history, tumor number, prognostic nutritional index, log (alpha-fetoprotein), and log (platelet-to-lymphocyte ratio). We named this final model the "multivariate prognostic model for Lenvatinib" (MPML), and a nomogram was constructed to predict the probability of survival at 6, 9, and 12 months. The MPML had good discrimination, calibration, and applicability. Cross-validation showed mean AUC values of 0.7779, 0.7738, and 0.7871 at 6, 9, and 12 months, respectively. According to nomogram points, mOS time was 21.57, 8.70, and 5.37 months in the low, medium, and high-risk groups, respectively (P < .001), and these differences were also observed in the PFS survival curve (P < .001).

CONCLUSIONS

The MPML stratified patients according to baseline clinical characteristics had a strong performance in predicting Lenvatinib efficacy and has the potential for use as an auxiliary clinical tool for individualized decision-making.

Prognostic model of immune checkpoint inhibitors combined with anti-angiogenic agents in unresectable hepatocellular carcinoma

Background

The combination of immune checkpoint inhibitors (ICIs) and anti-angiogenic agents has shown promising efficacy in unresectable hepatocellular carcinoma (HCC), but until now no clinical prognostic models or predictive biomarkers have been established.

Methods

From 2016 to 2021, a total of 258 HCCs treated with ICIs and tyrosine kinase inhibitors (TKIs) were retrospectively enrolled, as the study cohort. Patients' baseline data was extracted by least absolute and shrinkage selection operator (LASSO) and Cox regression. Finally, a prognostic model in the form of nomogram was developed. Model performance was assessed in terms of discrimination, calibration, and clinical utility. A 5-fold cross-validation was used to evaluate the internal repeatability of the model. In addition, the patient cohort was divided into three subgroups according to nomogram scores. Their survivals were estimated by Kaplan-Meier methods and the differences were analyzed using log-rank tests.

Results

Seven clinical parameters were selected: Eastern Cooperative Oncology Group performance status (ECOG PS), combination of transarterial chemoembolization (TACE), extrahepatic metastasis (EHM), platelet to lymphocyte ratio (PLR), alanine aminotransferase (ALT), alpha-fetoprotein (AFP), and Child-Pugh score. The model had an area under the curve (AUC) of 0.777 at 1 year and 0.772 at 2 years. Receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA) showed that the discrimination, consistency and applicability of the model were good. In addition, cross-validation validated the discrimination of the model, and the C index value of the model is 0.7405. The median overall survival (OS) of the high-, medium- and low-risk subgroups was 7.58, 17.50 and 53.17 months, respectively, with a significant difference between the groups (P < 0.0001).

Conclusion

We developed a comprehensive and simple prognostic model for the combination of ICIs plus TKIs. And it may predict the efficacy of the combination regimen for unresectable HCC.

Prognostic models for outcome prediction in patients with advanced hepatocellular carcinoma treated by systemic therapy: a systematic review and critical appraisal

Objective

To describe and analyze the predictive models of the prognosis of patients with hepatocellular carcinoma (HCC) undergoing systemic treatment.

Design

Systematic review.

Data sources

PubMed and Embase until December 2020 and manually searched references from eligible articles.

Eligibility criteria for study selection

The development, validation, or updating of prognostic models of patients with HCC after systemic treatment.

Results

The systematic search yielded 42 eligible articles: 28 articles described the development of 28 prognostic models of patients with HCC treated with systemic therapy, and 14 articles described the external validation of 32 existing prognostic models of patients with HCC undergoing systemic treatment. Among the 28 prognostic models, six were developed based on genes, of which five were expressed in full equations; the other 22 prognostic models were developed based on common clinical factors. Of the 28 prognostic models, 11 were validated both internally and externally, nine were validated only internally, two were validated only externally, and the remaining six models did not undergo any type of validation. Among the 28 prognostic models, the most common systemic treatment was sorafenib (n = 19); the most prevalent endpoint was overall survival (n = 28); and the most commonly used predictors were alpha-fetoprotein (n = 15), bilirubin (n = 8), albumin (n = 8), Child–Pugh score (n = 8), extrahepatic metastasis (n = 7), and tumor size (n = 7). Further, among 32 externally validated prognostic models, 12 were externally validated > 3 times.

Conclusions

This study describes and analyzes the prognostic models developed and validated for patients with HCC who have undergone systemic treatment. The results show that there are some methodological flaws in the model development process, and that external validation is rarely performed. Future research should focus on validating and updating existing models, and evaluating the effects of these models in clinical practice.

Systematic review registration

PROSPERO CRD42020200187.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-022-09841-5.

Early Predictive Value of Circulating Biomarkers for Sorafenib in Advanced Hepatocellular Carcinoma

INTRODUCTION

Sorafenib is currently the first-line therapeutic regimen for patients with advanced hepatocellular carcinoma (HCC). However, many patients did not experience any benefit and suffered extreme adverse events and heavy economic burden. Thus, the early identification of patients who are most likely to benefit from sorafenib is needed.

AREAS COVERED

This review focused on the clinical application of circulating biomarkers (including conventional biomarkers, immune biomarkers, genetic biomarkers, and some novel biomarkers) in advanced HCC patients treated with sorafenib. An online search on PubMed, Web of Science, Embase, and Cochrane Library was conducted from the inception to Aug 15, 2021. Studies investigating the predictive or prognostic value of these biomarkers were included.

EXPERT OPINION

The distinction of patients who may benefit from sorafenib treatment is of utmost importance. The predictive roles of circulating biomarkers could solve this problem. Many biomarkers can be obtained by liquid biopsy, which is a less or non-invasive approach. The short half-life of sorafenib could reflect the dynamic changes of tumor progression and monitor the treatment response. Circulating biomarkers obtained from liquid biopsy resulted as a promising assessment method in HCC, allowing for better treatment decisions in the near future.

The Relationship between albumin-Bilirubin grade and survival in hepatocelluler carcinoma patients treated with sorefanib

Background

Hepatocellular carcinoma (HCC) is one of the most common hepatic malignancies and the sixth most common cancer worldwide. Prognosis is affected by tumor stage, hepatic disfunction and patient performance. Albumin - bilirubin grade was developed to assess the hepatic function in patient with HCC. Aims: The purpose of this study was to examine the relationship between albumin-bilirubin (ALBI) grade and survival in HCC patients receiving sorafenib. We also planned to investigate whether ALBI scores in advanced stage patients are prognostic and predictive.

Patients and Methods

Patients presenting to the Karadeniz Technical University Medical Faculty Medical Oncology Clinic and diagnosed with HCC in 2010-2018 were included in the study. Fifty-six patients using sorafenib with Eastern Cooperative Oncology Group (ECOG) performance scores of 0, 1, or 2, who had not previously received systemic therapy were enrolled.

Results

Patients' median age was 64.8 years (range: 23-86), and 80.4% were men. The highest proportion of patients were infected with hepatitis B virus (46.4%), 37 patients were ECOG 1 (66.1%), and 40 were ALBI grade 2 (71.4%). The change occurring in ALBI scores after sorafenib therapy compared to pre-sorafenib values was found to affect progression-free survival. Prognosis was better in the group with decreasing ALBI scores than in the increasing score group (p: 0.028). Multivariate regression analysis revealed that the change occurring in ALBI scores after sorafenib therapy compared to pre-sorafenib values was predictive of progression-free survival independently of alpha-fetoprotein (AFP) levels.

Conclusion

This study shows that ALBI grade affects survival independently of AFP, Hand-Foot Syndrome (HFS), and other prognostic factors. ALBI grading can be used as a prognostic parameter in patients using sorafenib.

Usefulness of neutrophil-to-lymphocyte ratio (NLR) as a prognostic predictor after treatment of hepatocellular carcinoma." Review article

The neutrophil-to-lymphocyte ratio (NLR) is an inflammatory marker which has been investigated as a prognostic indicator in post-therapeutic recurrence and survival of patients with HCC. Our aim was to review all studies that assessed the prognostic value of pre-treatment NLR in predicting patient survival, cancer recurrence, and graft survival in patients undergoing various therapies for HCC. We searched the database of PubMed and Google Scholar to review all studies that have the word "NLR" and the word "HCC." We included all studies that assessed pre-treatment NLR as a prognostic factor in predicting outcomes in HCC patients. We excluded studies that assessed the correlation between post-treatment NLR or dynamic changes in NLR after treatment and HCC outcomes in an effort to minimize the confounding effect of each treatment on NLR. We reviewed 123 studies that studied the correlation between pre-treatment NLR and patient survival, 72 studies that evaluated the correlation between pre-treatment NLR and tumor recurrence, 21 studies that evaluated the correlation between NLR and tumor behavior, and 4 studies that assessed the correlation between NLR and graft survival. We found a remarkable heterogeneity between the methods of the studies, which is likely responsible for the differences in outcomes. The majority of the studies suggested a correlation between higher levels of pre-treatment NLR and poor outcomes. We concluded that NLR is a reliable and inexpensive biomarker and should be incorporated into other prognostic models to help determine outcomes following HCC treatment.

The importance of the glucose-to-lymphocyte ratio in patients with hepatocellular carcinoma treated with sorafenib

Aim: To show the prognostic significance of the glucose-to-lymphocyte ratio (GLR) in hepatocellular carcinoma (HCC). Patients & methods: A total of 150 patients with advanced HCC who were treated with sorafenib in our center between January 2011 and December 2019 were included in the study retrospectively. Neutrophil-to-lymphocyte ratio, systemic immune-inflammation index, lymphocyte-to-monocyte ratio, platelet-to-lymphocyte ratio, prognostic nutritional index and GLR were analyzed to assess their prognostic value using Kaplan-Meier and Cox regression analysis before and after propensity score matching (PSM). Results: In univariate analysis before and after PSM, albumin-bilirubin grade, neutrophil-to-lymphocyte ratio, systemic immune-inflammation index, lymphocyte-to-monocyte ratio, prognostic nutritional index, AFP level and GLR were found to be significantly associated with both progression-free and overall survival. In multivariate analysis before and after PSM, GLR, albumin-bilirubin grade and AFP were determined to be independent prognostic factors for progression-free and overall survival. Conclusion: The GLR prior to sorafenib treatment is a new prognostic biomarker that may predict survival in advanced HCC.

Role of the prognostic nutritional index in predicting survival in advanced hepatocellular carcinoma treated with regorafenib

AIM

A link has been established between malnutrition, immunological status, and hepatocellular carcinoma (HCC). The prognostic nutritional index (PNI) has been recognized as a prognostic indicator in early-stage HCC and in patients treated with first-line therapy. However, to date, the role of the PNI in HCC patients treated with regorafenib has not been reported.

METHODS

We undertook a multicentric analysis on a cohort of 284 patients affected by advanced HCC treated with regorafenib. The PNI was calculated as follows: 10 × serum albumin concentration (g/dl) + 0.005 × peripheral lymphocyte count (number/mm³ ). Univariate and multivariate analyses were used to investigate the association between PNI and survival outcomes.

RESULTS

A PNI cut-off value of 44.45 was calculated by a receiver operating characteristic analysis. The median overall survival was 12.8 and 7.8 months for patients with high (>44.45) and low (≤44.45) PNI, respectively (hazard ratio, 0.58; 95% confidence interval, 0.43-0.77; p = 0.0002). In the univariate and multivariate analyses, low PNI value and increased serum bilirubin level emerged as independent prognostic factors for overall survival. No differences were found between high and low PNI in terms of progression-free survival (p = 0.14).

CONCLUSION

If validated, the PNI could represent an easy-to-use prognostic tool able to guide the clinical decision-making process in HCC patients treated with regorafenib.

Carcinogenesis: Failure of resolution of inflammation?

Inflammation in the tumor microenvironment is a hallmark of cancer and is recognized as a key characteristic of carcinogens. However, the failure of resolution of inflammation in cancer is only recently being understood. Products of arachidonic acid and related fatty acid metabolism called eicosanoids, including prostaglandins, leukotrienes, lipoxins, and epoxyeicosanoids, critically regulate inflammation, as well as its resolution. The resolution of inflammation is now appreciated to be an active biochemical process regulated by endogenous specialized pro-resolving lipid autacoid mediators which combat infections and stimulate tissue repair/regeneration. Environmental and chemical human carcinogens, including aflatoxins, asbestos, nitrosamines, alcohol, and tobacco, induce tumor-promoting inflammation and can disrupt the resolution of inflammation contributing to a devastating global cancer burden. While mechanisms of carcinogenesis have focused on genotoxic activity to induce mutations, nongenotoxic mechanisms such as inflammation and oxidative stress promote genotoxicity, proliferation, and mutations. Moreover, carcinogens initiate oxidative stress to synergize with inflammation and DNA damage to fuel a vicious feedback loop of cell death, tissue damage, and carcinogenesis. In contrast, stimulation of resolution of inflammation may prevent carcinogenesis by clearance of cellular debris via macrophage phagocytosis and inhibition of an eicosanoid/cytokine storm of pro-inflammatory mediators. Controlling the host inflammatory response and its resolution in carcinogen-induced cancers will be critical to reducing carcinogen-induced morbidity and mortality. Here we review the recent evidence that stimulation of resolution of inflammation, including pro-resolution lipid mediators and soluble epoxide hydrolase inhibitors, may be a new chemopreventive approach to prevent carcinogen-induced cancer that should be evaluated in humans.

Pretreatment Neutrophil-to-Lymphocyte Ratio Predicts Survival and Liver Toxicity in Patients With Hepatocellular Carcinoma Treated With Stereotactic Ablative Radiation Therapy

PURPOSE

The objective of this study was to determine whether pretreatment neutrophil-to-lymphocyte ratio (NLR) could predict survival outcomes and liver toxicity in hepatocellular carcinoma (HCC) patients treated with stereotactic ablative radiation therapy (SABR).

METHODS AND MATERIALS

In this retrospective study we collected pretreatment NLR of HCC patients treated with SABR between December 2007 and August 2018 and determined its association with overall survival (OS), progression-free survival, and radiation-related liver toxicity defined as an increase in the Child-Turcotte-Pugh score by ≥2 within 3 months after SABR in the absence of disease progression.

RESULTS

A total of 153 patients with a median follow-up of 13.3 months were included. Receiver operating characteristic curve analysis found that an NLR ≥2.4 was optimum (area under the curve, 0.762; 95% confidence interval [CI], 0.682-0.841, P < .001) for predicting poor 1-year OS (38.2% vs 83.6%, P < .001). Multivariable analysis demonstrated that NLR was significantly associated with OS, both as a continuous (P = .006) and a binary variable (NLR set at 2.4; P = .003). Multiple tumors (P = .003), macrovascular invasion (P = .024), extrahepatic spread (P = .002), and albumin-bilirubin score (P = .020) were also significant predictors of OS. Elevated NLR independently prognosticated poor progression-free survival (P = .016). Liver toxicity was seen in 22 evaluable patients (15.4%). Receiver operating characteristic curve analysis found NLR ≥4.0 was optimum at predicting liver toxicity (31.4% vs 10.2%, P = .005). A higher NLR (P = .049) and albumin-bilirubin score (P = .002) were independent risk factors for liver toxicity.

CONCLUSIONS

NLR is an objective and ubiquitous inflammatory marker that can predict OS and liver toxicity in HCC patients undergoing SABR. NLR could be a useful biomarker for patient risk stratification and therapeutic decision-making.

The Key Characteristics of Carcinogens: Relationship to the Hallmarks of Cancer, Relevant Biomarkers, and Assays to Measure Them

The key characteristics (KC) of human carcinogens provide a uniform approach to evaluating mechanistic evidence in cancer hazard identification. Refinements to the approach were requested by organizations and individuals applying the KCs. We assembled an expert committee with knowledge of carcinogenesis and experience in applying the KCs in cancer hazard identification. We leveraged this expertise and examined the literature to more clearly describe each KC, identify current and emerging assays and in vivo biomarkers that can be used to measure them, and make recommendations for future assay development. We found that the KCs are clearly distinct from the Hallmarks of Cancer, that interrelationships among the KCs can be leveraged to strengthen the KC approach (and an understanding of environmental carcinogenesis), and that the KC approach is applicable to the systematic evaluation of a broad range of potential cancer hazards in vivo and in vitro We identified gaps in coverage of the KCs by current assays. Future efforts should expand the breadth, specificity, and sensitivity of validated assays and biomarkers that can measure the 10 KCs. Refinement of the KC approach will enhance and accelerate carcinogen identification, a first step in cancer prevention.See all articles in this CEBP Focus section, "Environmental Carcinogenesis: Pathways to Prevention."

Association does not imply prediction: the accuracy of birthweight in predicting child mortality and anthropometric failure

PURPOSE

Epidemiologic studies often conflate the strength of association with predictive accuracy and build classification models based on arbitrarily selected probability cutoffs without considering the cost of misclassification. We illustrated these common pitfalls by building association, prediction, and classification models using birthweight as an exposure and child mortality and child anthropometric failure as outcomes.

METHODS

Nationally representative samples of 188,819 and 164,113 children aged less than 5 years across India were used for our analysis of mortality and anthropometric failure, respectively. We assessed outcomes of neonatal, postneonatal, and child mortality as well as stunting, wasting, and underweight. Birthweight was the main exposure. We used adjusted and unadjusted logistic regression models to evaluate association strength, univariable and multivariable logistic regression models trained on 80% of the data using 10-fold cross-validation to evaluate predictive power, and classification models across a series of possible misclassification cost scenarios to evaluate classification accuracy.

RESULTS

Birthweight was strongly associated with five of six outcomes (P < .001), and associations were robust to covariate adjustment. Prediction models evaluated on the test set showed that birthweight was a poor discriminator of all outcomes (area under the curve < 0.62), and that adding birthweight to a multivariable model did not meaningfully improve discrimination. Prediction models for anthropometric failure showed substantially better calibration than prediction models for mortality. Depending on the ratio of false positive (FP) cost to false negative (FN) cost, the probability cutoff that minimized total misclassification cost ranged from 0.116 (cost ratio = 7:93) to 0.706 (cost ratio = 4:1), TPR ranged from 0.999 to 0.004, and PPV ranged from 0.355 to 0.867..

CONCLUSIONS

Although birthweight is strongly associated with mortality and anthropometric failure, it is a poor predictor of child health outcomes, highlighting that strong associations do not imply predictive power. We recommend that (1) future research focus on building predictive models for anthropometric failure given their clinical relevance in diagnosing individual cases, and that (2) studies that build classifiers report performance metrics across a range of cutoffs to account for variation in the cost of FPs and FNs.

Prognostic Roles of Blood Inflammatory Markers in Hepatocellular Carcinoma Patients Taking Sorafenib. A Systematic Review and Meta-Analysis

Objective: The purpose of this meta-analysis is to investigate the effectiveness of the prognostic roles of blood inflammatory markers in hepatocellular carcinoma (HCC) patients receiving sorafenib.

Methods: We carried out a comprehensive literature search in four databases. Study endpoints, hazard ratios (HRs) and the associated 95% confidence intervals (CI) for clinical outcomes, which were to assess therapeutic efficacy, were extracted. This meta-analysis was conducted by Review Manager 5.3.

Results: We summarized the available evidence from 18 studies with a total of 2,745 cases. The pooled results showed that the synthesized HR favored patients with low pretreatment NLR (neutrophil-to-lymphocyte ratio), which also indicated that HCC patients with a lower baseline NLR may have a better response to sorafenib than those with higher NLR (HR = 1.76, 95% CI [1.44, 2.15], P < 0.00001, I² = 68%). Significance was also observed for the prognostic function of the PLR (platelet-to-lymphocyte ratio) of HCC patients treated with sorafenib (HR = 1.49, 95% CI [1.16, 1.93], P = 0.002, I² = 0%, P = 0.65). The subgroup analysis revealed that different gene backgrounds play a prominent role in the source of heterogeneity. Interestingly, the predictive effect on OS (overall survival) was more pronounced as the NLR cutoff value increased. Notably, a significant predictive effect of NLR on the clinical outcome was detected in HCC patients treated with sorafenib compared to those treated with tivantinib.

Conclusion: In conclusion, the present study reported promising predictive biomarkers for HCC patients and notably indicated that HCC patients with a lower baseline NLR and PLR may have a better response to sorafenib than those with higher ones. Further large-scale prospective studies are required to determine the optimal NLR and PLR cutoff values, which are important for identifying the dominant populations for sorafenib treatment.

The role of PNI to predict survival in advanced hepatocellular carcinoma treated with Sorafenib

BACKGROUND AND AIMS

The present study aims to investigate the role of the prognostic nutritional index (PNI) on survival in patients with advanced hepatocellular carcinoma (HCC) treated with sorafenib.

METHODS

This multicentric study included a training cohort of 194 HCC patients and three external validation cohorts of 129, 76 and 265 HCC patients treated with Sorafenib, respectively. The PNI was calculated as follows: 10 × serum albumin (g/dL) + 0.005 × total lymphocyte count (per mm3). Univariate and multivariate analyses were performed to investigate the association between the covariates and the overall survival (OS).

RESULTS

A PNI cut-off value of 31.3 was established using the ROC analysis. In the training cohort, the median OS was 14.8 months (95% CI 12-76.3) and 6.8 months (95% CI 2.7-24.6) for patients with a high (>31.3) and low (<31.3) PNI, respectively. At both the univariate and the multivariate analysis, low PNI value (p = 0.0004), a 1-unit increase of aspartate aminotransferase (p = 0.0001), and age > 70 years (p< 0.0038) were independent prognostic factors for OS. By performing the same multivariate analysis of the training cohort, the PNI <31.3 versus >31.3 was found to be an independent prognostic factor for predicting OS in all the three validation cohorts.

CONCLUSIONS

PNI represents a prognostic tool in advanced HCC treated with first-line Sorafenib. It is readily available and low-cost, and it could be implemented in clinical practice in patients with HCC.

Three myths about risk thresholds for prediction models

BACKGROUND

Clinical prediction models are useful in estimating a patient's risk of having a certain disease or experiencing an event in the future based on their current characteristics. Defining an appropriate risk threshold to recommend intervention is a key challenge in bringing a risk prediction model to clinical application; such risk thresholds are often defined in an ad hoc way. This is problematic because tacitly assumed costs of false positive and false negative classifications may not be clinically sensible. For example, when choosing the risk threshold that maximizes the proportion of patients correctly classified, false positives and false negatives are assumed equally costly. Furthermore, small to moderate sample sizes may lead to unstable optimal thresholds, which requires a particularly cautious interpretation of results.

MAIN TEXT

We discuss how three common myths about risk thresholds often lead to inappropriate risk stratification of patients. First, we point out the contexts of counseling and shared decision-making in which a continuous risk estimate is more useful than risk stratification. Second, we argue that threshold selection should reflect the consequences of the decisions made following risk stratification. Third, we emphasize that there is usually no universally optimal threshold but rather that a plausible risk threshold depends on the clinical context. Consequently, we recommend to present results for multiple risk thresholds when developing or validating a prediction model.

CONCLUSION

Bearing in mind these three considerations can avoid inappropriate allocation (and non-allocation) of interventions. Using discriminating and well-calibrated models will generate better clinical outcomes if context-dependent thresholds are used.