Diagnostic Models Combining Clinical Information, Ultrasound and Biochemical Markers for Ovarian Cancer: Cochrane Systematic Review and Meta-Analysis.

Background: Ovarian cancer (OC) is a diagnostic challenge, with the majority diagnosed at late stages. Existing systematic reviews of diagnostic models either use inappropriate meta-analytic methods or do not conduct statistical comparisons of models or stratify test performance by menopausal status. Methods: We searched CENTRAL, MEDLINE, EMBASE, CINAHL, CDSR, DARE, Health Technology Assessment Database and SCI Science Citation Index, trials registers, conference proceedings from 1991 to June 2019. Cochrane collaboration review methods included QUADAS-2 quality assessment and meta-analysis using hierarchical modelling. RMI, ROMA or ADNEX at any test positivity threshold were investigated. Histology or clinical follow-up was the reference standard. We excluded screening studies, studies restricted to pregnancy, recurrent or metastatic OC. 2 × 2 diagnostic tables were extracted separately for pre- and post-menopausal women. Results: We included 58 studies (30,121 patients, 9061 cases of ovarian cancer). Prevalence of OC ranged from 16 to 55% in studies. For premenopausal women, ROMA at a threshold of 13.1 (+/−2) and ADNEX at a threshold of 10% demonstrated significantly higher sensitivity compared to RMI I at 200 (p < 0.0001) 77.8 (72.5, 82.4), 94.9 (92.5, 96.6), and 57.1% (50.6 to 63.4) but lower specificity (p < 0.002), 92.5 (90.0, 94.4), 84.3 (81.3, 86.8), and 78.2 (75.8, 80.4). For postmenopausal women, ROMA at a threshold of 27.7 (+/−2) and AdNEX at a threshold of 10% demonstrated significantly higher sensitivity compared to RMI I at a threshold of 200 (p < 0.001) 90.4 (87.4, 92.7), 97.6 (96.2, 98.5), and 78.7 (74.3, 82.5), specificity of ROMA was comparable, whilst ADneX was lower, 85.5 (81.3, 88.9), 81.3 (76.9, 85.0) (p = 0.155), compared to RMI 55.2 (51.2, 59.1) (p < 0.001). Conclusions: In pre-menopausal women, ROMA and ADNEX offer significantly higher sensitivity but significantly decreased specificity. In post-menopausal women, ROMA demonstrates significantly higher sensitivity and comparable specificity to RMI I, ADNEX has the highest sensitivity of all models, but with significantly reduced specificity. RMI I has poor sensitivity compared to ROMA or ADNEX. Choice between ROMA and ADNEX as a replacement test will depend on cost effectiveness and resource implications.

Menopausal status, ultrasound and biomarker tests in combination for the diagnosis of ovarian cancer in symptomatic women.

BACKGROUND: Ovarian cancer (OC) has the highest case fatality rate of all gynaecological cancers. Diagnostic delays are caused by non-specific symptoms. Existing systematic reviews have not comprehensively covered tests in current practice, not estimated accuracy separately in pre- and postmenopausal women, or used inappropriate meta-analytic methods. OBJECTIVES: To establish the accuracy of combinations of menopausal status, ultrasound scan (USS) and biomarkers for the diagnosis of ovarian cancer in pre- and postmenopausal women and compare the accuracy of different test combinations. SEARCH METHODS: We searched CENTRAL, MEDLINE (Ovid), Embase (Ovid), five other databases and three trial registries from 1991 to 2015 and MEDLINE (Ovid) and Embase (Ovid) form June 2015 to June 2019. We also searched conference proceedings from the European Society of Gynaecological Oncology, International Gynecologic Cancer Society, American Society of Clinical Oncology and Society of Gynecologic Oncology, ZETOC and Conference Proceedings Citation Index (Web of Knowledge). We searched reference lists of included studies and published systematic reviews. SELECTION CRITERIA: We included cross-sectional diagnostic test accuracy studies evaluating single tests or comparing two or more tests, randomised trials comparing two or more tests, and studies validating multivariable models for the diagnosis of OC investigating test combinations, compared with a reference standard of histological confirmation or clinical follow-up in women with a pelvic mass (detected clinically or through USS) suspicious for OC. DATA COLLECTION AND ANALYSIS: Two review authors independently extracted data and assessed quality using QUADAS-2. We used the bivariate hierarchical model to indirectly compare tests at commonly reported thresholds in pre- and postmenopausal women separately. We indirectly compared tests across all thresholds and estimated sensitivity at fixed specificities of 80% and 90% by fitting hierarchical summary receiver operating characteristic (HSROC) models in pre- and postmenopausal women separately. MAIN RESULTS: We included 59 studies (32,059 women, 9545 cases of OC). Two tests evaluated the accuracy of a combination of menopausal status and USS findings (IOTA Logistic Regression Model 2 (LR2) and the Assessment of Different NEoplasias in the adneXa model (ADNEX)); one test evaluated the accuracy of a combination of menopausal status, USS findings and serum biomarker CA125 (Risk of Malignancy Index (RMI)); and one test evaluated the accuracy of a combination of menopausal status and two serum biomarkers (CA125 and HE4) (Risk of Ovarian Malignancy Algorithm (ROMA)). Most studies were at high or unclear risk of bias in participant, reference standard, and flow and timing domains. All studies were in hospital settings. Prevalence was 16% (RMI, ROMA), 22% (LR2) and 27% (ADNEX) in premenopausal women and 38% (RMI), 45% (ROMA), 52% (LR2) and 55% (ADNEX) in postmenopausal women. The prevalence of OC in the studies was considerably higher than would be expected in symptomatic women presenting in community-based settings, or in women referred from the community to hospital with a suspicion of OC. Studies were at high or unclear applicability because presenting features were not reported, or USS was performed by experienced ultrasonographers for RMI, LR2 and ADNEX. The higher sensitivity and lower specificity observed in postmenopausal compared to premenopausal women across all index tests and at all thresholds may reflect highly selected patient cohorts in the included studies. In premenopausal women, ROMA at a threshold of 13.1 (± 2), LR2 at a threshold to achieve a post-test probability of OC of 10% and ADNEX (post-test probability 10%) demonstrated a higher sensitivity (ROMA: 77.4%, 95% CI 72.7% to 81.5%; LR2: 83.3%, 95% CI 74.7% to 89.5%; ADNEX: 95.5%, 95% CI 91.0% to 97.8%) compared to RMI (57.2%, 95% CI 50.3% to 63.8%). The specificity of ROMA and ADNEX were lower in premenopausal women (ROMA: 84.3%, 95% CI 81.2% to 87.0%; ADNEX: 77.8%, 95% CI 67.4% to 85.5%) compared to RMI 92.5% (95% CI 90.3% to 94.2%). The specificity of LR2 was comparable to RMI (90.4%, 95% CI 84.6% to 94.1%). In postmenopausal women, ROMA at a threshold of 27.7 (± 2), LR2 (post-test probability 10%) and ADNEX (post-test probability 10%) demonstrated a higher sensitivity (ROMA: 90.3%, 95% CI 87.5% to 92.6%; LR2: 94.8%, 95% CI 92.3% to 96.6%; ADNEX: 97.6%, 95% CI 95.6% to 98.7%) compared to RMI (78.4%, 95% CI 74.6% to 81.7%). Specificity of ROMA at a threshold of 27.7 (± 2) (81.5, 95% CI 76.5% to 85.5%) was comparable to RMI (85.4%, 95% CI 82.0% to 88.2%), whereas for LR2 (post-test probability 10%) and ADNEX (post-test probability 10%) specificity was lower (LR2: 60.6%, 95% CI 50.5% to 69.9%; ADNEX: 55.0%, 95% CI 42.8% to 66.6%). AUTHORS' CONCLUSIONS: In specialist healthcare settings in both premenopausal and postmenopausal women, RMI has poor sensitivity. In premenopausal women, ROMA, LR2 and ADNEX offer better sensitivity (fewer missed cancers), but for ROMA and ADNEX this is off-set by a decrease in specificity and increase in false positives. In postmenopausal women, ROMA demonstrates a higher sensitivity and comparable specificity to RMI. ADNEX has the highest sensitivity in postmenopausal women, but reduced specificity. The prevalence of OC in included studies is representative of a highly selected referred population, rather than a population in whom referral is being considered. The comparative accuracy of tests observed here may not be transferable to non-specialist settings. Ultimately health systems need to balance accuracy and resource implications to identify the most suitable test.

Guaiac-based faecal occult blood tests versus faecal immunochemical tests for colorectal cancer screening in average-risk individuals.

BACKGROUND: Worldwide, many countries have adopted colorectal cancer (CRC) screening programmes, often based on faecal occult blood tests (FOBTs). CRC screening aims to detect advanced neoplasia (AN), which is defined as CRC or advanced adenomas. FOBTs fall into two categories based on detection technique and the detected blood component: qualitative guaiac-based FOBTs (gFOBTs) and faecal immunochemical tests (FITs), which can be qualitative and quantitative. Screening with gFOBTs reduces CRC-related mortality. OBJECTIVES: To compare the diagnostic test accuracy of gFOBT and FIT screening for detecting advanced colorectal neoplasia in average-risk individuals. SEARCH METHODS: We searched CENTRAL, MEDLINE, Embase, BIOSIS Citation Index, Science Citation Index Expanded, and Google Scholar. We searched the reference lists and PubMed-related articles of included studies to identify additional studies. SELECTION CRITERIA: We included prospective and retrospective studies that provided the number of true positives, false positives, false negatives, and true negatives for gFOBTs, FITs, or both, with colonoscopy as reference standard. We excluded case-control studies. We included studies in which all participants underwent both index test and reference standard ("reference standard: all"), and studies in which only participants with a positive index test underwent the reference standard while participants with a negative test were followed for at least one year for development of interval carcinomas ("reference standard: positive"). The target population consisted of asymptomatic, average-risk individuals undergoing CRC screening. The target conditions were CRC and advanced neoplasia (advanced adenomas and CRC combined). DATA COLLECTION AND ANALYSIS: Two review authors independently screened and selected studies for inclusion. In case of disagreement, a third review author made the final decision. We used the Rutter and Gatsonis hierarchical summary receiver operating characteristic model to explore differences between tests and identify potential sources of heterogeneity, and the bivariate hierarchical model to estimate sensitivity and specificity at common thresholds: 10 µg haemoglobin (Hb)/g faeces and 20 µg Hb/g faeces. We performed indirect comparisons of the accuracy of the two tests and direct comparisons when both index tests were evaluated in the same population. MAIN RESULTS: We ran the initial search on 25 June 2019, which yielded 63 studies for inclusion. We ran a top-up search on 14 September 2021, which yielded one potentially eligible study, currently awaiting classification. We included a total of 33 "reference standard: all" published articles involving 104,640 participants. Six studies evaluated only gFOBTs, 23 studies evaluated only FITs, and four studies included both gFOBTs and FITs. The cut-off for positivity of FITs varied between 2.4 µg and 50 µg Hb/g faeces. For each Quality Assessment of Diagnostic Accuracy Studies (QUADAS)-2 domain, we assessed risk of bias as high in less than 20% of studies. The summary curve showed that FITs had a higher discriminative ability than gFOBTs for AN (P < 0.001) and CRC (P = 0.004). For the detection of AN, the summary sensitivity of gFOBTs was 15% (95% confidence interval (CI) 12% to 20%), which was significantly lower than FITs at both 10 µg and 20 µg Hb/g cut-offs with summary sensitivities of 33% (95% CI 27% to 40%; P < 0.001) and 26% (95% CI 21% to 31%, P = 0.002), respectively. Results were simulated in a hypothetical cohort of 10,000 screening participants with 1% CRC prevalence and 10% AN prevalence. Out of 1000 participants with AN, gFOBTs missed 850, while FITs missed 670 (10 µg Hb/g cut-off) and 740 (20 µg Hb/g cut-off). No significant differences in summary specificity for AN detection were found between gFOBTs (94%; 95% CI 92% to 96%), and FITs at 10 µg Hb/g cut-off (93%; 95% CI 90% to 95%) and at 20 µg Hb/g cut-off (97%; 95% CI 95% to 98%). So, among 9000 participants without AN, 540 were offered (unnecessary) colonoscopy with gFOBTs compared to 630 (10 µg Hb/g) and 270 (20 µg Hb/g) with FITs. Similarly, for the detection of CRC, the summary sensitivity of gFOBTs, 39% (95% CI 25% to 55%), was significantly lower than FITs at 10 µg and 20 µg Hb/g cut-offs: 76% (95% CI 57% to 88%: P = 0.001) and 65% (95% CI 46% to 80%; P = 0.035), respectively. So, out of 100 participants with CRC, gFOBTs missed 61, and FITs missed 24 (10 µg Hb/g) and 35 (20 µg Hb/g). No significant differences in summary specificity for CRC were found between gFOBTs (94%; 95% CI 91% to 96%), and FITs at the 10 µg Hb/g cut-off (94%; 95% CI 87% to 97%) and 20 µg Hb/g cut-off (96%; 95% CI 91% to 98%). So, out of 9900 participants without CRC, 594 were offered (unnecessary) colonoscopy with gFOBTs versus 594 (10 µg Hb/g) and 396 (20 µg Hb/g) with FITs. In five studies that compared FITs and gFOBTs in the same population, FITs showed a higher discriminative ability for AN than gFOBTs (P = 0.003). We included a total of 30 "reference standard: positive" studies involving 3,664,934 participants. Of these, eight were gFOBT-only studies, 18 were FIT-only studies, and four studies combined both gFOBTs and FITs. The cut-off for positivity of FITs varied between 5 µg to 250 µg Hb/g faeces. For each QUADAS-2 domain, we assessed risk of bias as high in less than 20% of studies. The summary curve showed that FITs had a higher discriminative ability for detecting CRC than gFOBTs (P < 0.001). The summary sensitivity for CRC of gFOBTs, 59% (95% CI 55% to 64%), was significantly lower than FITs at the 10 µg Hb/g cut-off, 89% (95% CI 80% to 95%; P < 0.001) and the 20 µg Hb/g cut-off, 89% (95% CI 85% to 92%; P < 0.001). So, in the hypothetical cohort with 100 participants with CRC, gFOBTs missed 41, while FITs missed 11 (10 µg Hb/g) and 11 (20 µg Hb/g). The summary specificity of gFOBTs was 98% (95% CI 98% to 99%), which was higher than FITs at both 10 µg and 20 µg Hb/g cut-offs: 94% (95% CI 92% to 95%; P < 0.001) and 95% (95% CI 94% to 96%; P < 0.001), respectively. So, out of 9900 participants without CRC, 198 were offered (unnecessary) colonoscopy with gFOBTs compared to 594 (10 µg Hb/g) and 495 (20 µg Hb/g) with FITs. At a specificity of 90% and 95%, FITs had a higher sensitivity than gFOBTs. AUTHORS' CONCLUSIONS: FITs are superior to gFOBTs in detecting AN and CRC in average-risk individuals. Specificity of both tests was similar in "reference standard: all" studies, whereas specificity was significantly higher for gFOBTs than FITs in "reference standard: positive" studies. However, at pre-specified specificities, the sensitivity of FITs was significantly higher than gFOBTs.

The Impact of Diabetes and Glucose-Lowering Therapies on Hepatocellular Carcinoma Incidence and Overall Survival.

PURPOSE: The incidence of hepatocellular carcinoma (HCC) in the United Kingdom has increased 60% in the past 10 years. The epidemics of obesity and type 2 diabetes are contributing factors. In this article, we examine the impact of diabetes and glucose-lowering treatments on HCC incidence and overall survival (OS). METHODS: Data from 1064 patients diagnosed with chronic liver disease (CLD) (n = 340) or HCC (n = 724) were collected from 2007 to 2012. Patients with HCC were followed up prospectively. Univariate and multivariate logistic regression determined HCC risk factors. Kaplan-Meier curves were used to examine survival and Cox proportional hazards analysis estimated hazard ratios (HRs) for death according to use of glucose-lowering therapies. FINDINGS: Diabetes prevalence was 39.6% and 10.6% within the HCC and CLD cohorts, respectively. The odds ratio for having HCC in patients with diabetes was 5.55 (P < 0.001). Univariate analysis found an increased association of HCC with age, sex, cirrhosis, hemochromatosis, alcohol abuse, diabetes, and Child's Pugh score. In multivariate analysis age, sex, cirrhosis, Child's Pugh score, diabetes status, and insulin use retained significance. Diabetes status did not significantly affect OS in HCC; however, in people with diabetes and HCC, metformin treatment was associated with improved OS (mean survival, 31 vs 24 months; P =0.016; HR for death = 0.75; P = 0.032). IMPLICATIONS: Diabetes is significantly associated with HCC in the United Kingdom. Metformin treatment is associated with improved OS after HCC diagnosis. Treatment of diabetes should be appropriately reviewed in high-risk populations, with specific consideration of the potential hepatoprotective effects of metformin in HCC.

The prognostic and diagnostic significance of the neutrophil-to-lymphocyte ratio in hepatocellular carcinoma: a prospective controlled study.

BACKGROUND: The neutrophil-lymphocyte ratio (NLR), a presumed measure of the balance between neutrophil-associated pro-tumour inflammation and lymphocyte-dependent antitumour immune function, has been suggested as a prognostic factor for several cancers, including hepatocellular carcinoma (HCC). METHODS: In this study, a prospectively accrued cohort of 781 patients (493 HCC and 288 chronic liver disease (CLD) without HCC) were followed-up for more than 6 years. NLR levels between HCC and CLD patients were compared, and the effect of baseline NLR on overall survival amongst HCC patients was assessed via multivariable Cox regression analysis. RESULTS: On entry into the study ('baseline'), there was no clinically significant difference in the NLR values between CLD and HCC patients. Amongst HCC patients, NLR levels closest to last visit/death were significantly higher compared to baseline. Multivariable Cox regression analysis showed that NLR was an independent prognostic factor, even after adjustment for the HCC stage. CONCLUSION: NLR is a significant independent factor influencing survival in HCC patients, hence offering an additional dimension in prognostic models.

Sorafenib is associated with a reduced rate of tumour growth and liver function deterioration in HCV-induced hepatocellular carcinoma.

BACKGROUND & AIMS: Sorafenib has been the standard of care for patients with advanced hepatocellular carcinoma and although immunotherapeutic approaches are now challenging this position, it retains an advantage in HCV-seropositive patients. We aimed to quantify the rate of tumour progression in patients receiving sorafenib and relate this figure to survival, both overall, and according to viral status. METHODS: Using serial data from an international clinical trial we applied a joint model to combine survival and progression over time in order to estimate the rate of tumour growth as assessed by tumour burden and serum alpha-fetoprotein, and the impact of treatment on liver function. RESULTS: High tumour burden at baseline was associated with an increased risk of death. In patients still alive at the end of the study, the progression in relation to tumour burden was very low compared to those who died within the study. Overall, the change in mean tumour burden was 0.12 mm per day or an absolute growth rate of 3.6 mm/month. Median doubling time was 665 days. For those who progressed above 0.12 mm per day or the 12% rate, median survival was 234 days compared to 384 days if the rate was below 12%. Tumour growth rate and serum alpha-fetoprotein rise were significantly lower in those who were HCV seropositive as was the rate of decline in liver function. These results were replicated in 2 independent patient groups. CONCLUSION: Our analysis suggests that sorafenib treatment is associated with improved survival in patients with advanced hepatocellular carcinoma mainly by decreasing the rate of tumour growth and liver function deterioration among patients with HCV infection. LAY SUMMARY: Among patients receiving sorafenib for advanced hepatocellular carcinoma the rate of tumour growth (as assessed by changes in tumour size and the biomarker alpha-fetoprotein) and the deterioration of liver function is less in those who have the hepatitis C virus, than in those who do not.

Serum Levels of α-Fetoprotein Increased More Than 10 Years Before Detection of Hepatocellular Carcinoma.

BACKGROUND & AIMS: Ultrasound (US)-based screening has been recommended for patients with an increased risk of hepatocellular carcinoma (HCC). US analysis, however, is limited in patients who are obese or have small tumors. The addition of serum level of α-fetoprotein (AFP) measurements to US analysis can increase detection of HCC. We analyzed data from patients with chronic liver disease, collected over 15 years in an HCC surveillance program, to develop a model to assess risk of HCC. METHODS: We collected data from 3450 patients with chronic liver disease undergoing US surveillance in Japan from March 1998 through April 2014, and followed them up for a median of 8.83 years. We performed longitudinal discriminant analysis of serial AFP measurements (median number of observations/patient, 56; approximately every 3 months) to develop a model to determine the risk of HCC. We validated the model using data from 2 cohorts of patients with chronic liver disease in Japan (404 and 2754 patients) and 1 cohort in Scotland (1596 patients). RESULTS: HCC was detected in 413 patients (median tumor diameter, 1.8 cm), during a median follow-up time of 6.60 years. In the development data set, the model identified patients who developed HCC with an area under the curve of 0.78; it correctly identified 74.3% of patients who did develop HCC, and 72.9% of patients who did not. Overall, 73.1% of patients were classified correctly. The model could be used to assign patients to a high-risk group (27.5 HCCs/1000 patient-years) vs a low-risk group (4.9 HCCs/1000 patient-years). A similar performance was observed when the model was used to assess patients with cirrhosis. Analysis of the validation cohorts produced similar results. CONCLUSIONS: We developed and validated a model to identify patients with chronic liver disease who are at risk for HCC based on change in serum AFP level over time. The model could be used to assign patients to high-risk vs low-risk groups, and might be used to select patients for surveillance.

Improved survival prediction and comparison of prognostic models for patients with hepatocellular carcinoma treated with sorafenib.

BACKGROUND: The 'Prediction Of Survival in Advanced Sorafenib-treated HCC' (PROSASH) model addressed the heterogeneous survival of patients with hepatocellular carcinoma (HCC) treated with sorafenib in clinical trials but requires validation in daily clinical practice. This study aimed to validate, compare and optimize this model for survival prediction. METHODS: Patients treated with sorafenib for HCC at five tertiary European centres were retrospectively staged according to the PROSASH model. In addition, the optimized PROSASH-II model was developed using the data of four centres (training set) and tested in an independent dataset. These models for overall survival (OS) were then compared with existing prognostic models. RESULTS: The PROSASH model was validated in 445 patients, showing clear differences between the four risk groups (OS 16.9-4.6 months). A total of 920 patients (n = 615 in training set, n = 305 in validation set) were available to develop PROSASH-II. This optimized model incorporated fewer and less subjective parameters: the serum albumin, bilirubin and alpha-foetoprotein, and macrovascular invasion, extrahepatic spread and largest tumour size on imaging. Both PROSASH and PROSASH-II showed improved discrimination (C-index 0.62 and 0.63, respectively) compared with existing prognostic scores (C-index ≤0.59). CONCLUSIONS: In HCC patients treated with sorafenib, individualized prediction of survival and risk group stratification using baseline prognostic and predictive parameters with the PROSASH model was validated. The refined PROSASH-II model performed at least as good with fewer and more objective parameters. PROSASH-II can be used as a tool for tailored treatment of HCC in daily practice and to define pre-planned subgroups for future studies.

GALAD Score Detects Early Hepatocellular Carcinoma in an International Cohort of Patients With Nonalcoholic Steatohepatitis.

BACKGROUND & AIMS: The prevalence of nonalcoholic steatohepatitis (NASH) associated hepatocellular carcinoma (HCC) is increasing. However, strategies for detection of early-stage HCC in patients with NASH have limitations. We assessed the ability of the GALAD score, which determines risk of HCC based on patient sex; age; and serum levels of α-fetoprotein (AFP), AFP isoform L3 (AFP-L3), and des-gamma-carboxy prothrombin (DCP), to detect HCC in patients with NASH. METHODS: We performed a case-control study of 125 patients with HCC (20% within Milan Criteria) and 231 patients without HCC (NASH controls) from 8 centers in Germany. We compared the performance of serum AFP, AFP-L3, or DCP vs GALAD score to identify patients with HCC using receiver operating characteristic curves and corresponding area under the curve (AUC) analyses. We also analyzed data from 389 patients with NASH under surveillance for HCC in Japan, followed for a median of 167 months. During the 5-year screening period, 26 patients developed HCC. To compensate for irregular intervals of data points, we performed locally weighted scatterplot smoothing, linear regression, and a non-linear curve fit to assess development of GALAD before HCC development. RESULTS: The GALAD score identified patients with any stage HCC with an AUC of 0.96 - significantly greater than values for serum levels of AFP (AUC, 0.88), AFP-L3 (AUC, 0.86) or DCP (AUC, 0.87). AUC values for the GALAD score were consistent in patients with cirrhosis (AUC, 0.93) and without cirrhosis (AUC, 0.98). For detection of HCC within Milan Criteria, the GALAD score achieved an AUC of 0.91, with a sensitivity of 68% and specificity of 95% at a cutoff of -0.63. In a pilot Japanese cohort study, the mean GALAD score was higher in patients with NASH who developed HCC than in those who did not develop HCC as early as 1.5 years before HCC diagnosis. GALAD scores were above -0.63 approximately 200 days before the diagnosis of HCC. CONCLUSIONS: In a case-control study performed in Germany and a pilot cohort study in Japan, we found the GALAD score may detect HCC with high levels of accuracy in patients with NASH, with and without cirrhosis. The GALAD score can detect patients with early-stage HCC, and might facilitate surveillance of patients with NASH, who are often obese, which limits the sensitivity of detection of liver cancer by ultrasound.

The chances of hepatic resection curing hepatocellular carcinoma.

BACKGROUND & AIMS: The popular sense of the word "cure" implies that a patient treated for a specific disease will return to have the same life expectancy as if he/she had never had the disease. In analytic terms, it translates into the concept of statistical cure which occurs when a group of patients returns to having similar mortality to a reference population. The aim of this study was to assess the probability of being cured from hepatocellular carcinoma (HCC) by hepatic resection. METHODS: Data from 2,523 patients undergoing resection for HCC were used to fit statistical cure models, to compare disease-free survival (DFS) after surgery to the survival expected for patients with chronic hepatitis and/or cirrhosis and the general population, matched by sex, age, race/ethnicity and year of diagnosis. RESULTS: The probability of resection enabling patients with HCC to achieve the same life expectancy as those with chronic hepatitis and/or cirrhosis was 26.3%. The conditional probability of achieving this result was time-dependent, requiring about 8.9 years to be accomplished with 95% certainty. Considering the general population as a reference, the cure fraction decreased to 17.1%. Uncured patients had a median DFS of 1.5 years. In multivariable analysis, patient's age and the risk of early HCC recurrence (within 2 years) were independent determinants of the chance of cure (p <0.001). The chances of being cured ranged between 36.0% for individuals at low risk of early recurrence to approximately 3.6% for those at high risk. CONCLUSION: Estimates of the chance of being cured of HCC by resection showed that cure is achievable, and its likelihood increases with the passing of recurrence-free time. The data presented herein can be used to inform decision making and to provide patients with accurate information. LAY SUMMARY: Data from 2,523 patients who underwent resection for hepatocellular carcinoma were used to estimate the probability that resection would enable treated patients to achieve the same life expectancy as patients with chronic hepatitis and/or cirrhosis, and the general population. Herein, the cure model suggests that in patients with hepatocellular carcinoma, resection can enable patients to achieve the same life expectancy as those with chronic liver disease in 26.3% of cases and as the general population in 17.1% of cases.

Prediction of Survival Among Patients Receiving Transarterial Chemoembolization for Hepatocellular Carcinoma: A Response-Based Approach.

BACKGROUND AND AIMS: The heterogeneity of intermediate-stage hepatocellular carcinoma (HCC) and the widespread use of transarterial chemoembolization (TACE) outside recommended guidelines have encouraged the development of scoring systems that predict patient survival. The aim of this study was to build and validate statistical models that offer individualized patient survival prediction using response to TACE as a variable. APPROACH AND RESULTS: Clinically relevant baseline parameters were collected for 4,621 patients with HCC treated with TACE at 19 centers in 11 countries. In some of the centers, radiological responses (as assessed by modified Response Evaluation Criteria in Solid Tumors [mRECIST]) were also accrued. The data set was divided into a training set, an internal validation set, and two external validation sets. A pre-TACE model ("Pre-TACE-Predict") and a post-TACE model ("Post-TACE-Predict") that included response were built. The performance of the models in predicting overall survival (OS) was compared with existing ones. The median OS was 19.9 months. The factors influencing survival were tumor number and size, alpha-fetoprotein, albumin, bilirubin, vascular invasion, cause, and response as assessed by mRECIST. The proposed models showed superior predictive accuracy compared with existing models (the hepatoma arterial embolization prognostic score and its various modifications) and allowed for patient stratification into four distinct risk categories whose median OS ranged from 7 months to more than 4 years. CONCLUSIONS: A TACE-specific and extensively validated model based on routinely available clinical features and response after first TACE permitted patient-level prognostication.

Using prognostic and predictive clinical features to make personalised survival prediction in advanced hepatocellular carcinoma patients undergoing sorafenib treatment.

BACKGROUND: Sorafenib is the current standard of care for patients with advanced hepatocellular carcinoma (aHCC) and has been shown to improve survival by about 3 months compared to placebo. However, survival varies widely from under three months to over two years. The aim of this study was to build a statistical model that allows personalised survival prediction following sorafenib treatment. METHODS: We had access to 1130 patients undergoing sorafenib treatment for aHCC as part of the control arm for two phase III randomised clinical trials (RCTs). A multivariable model was built that predicts survival based on baseline clinical features. The statistical approach permits both group-level risk stratification and individual-level survival prediction at any given time point. The model was calibrated, and its discrimination assessed through Harrell's c-index and Royston-Sauerbrei's R2D. RESULTS: The variables influencing overall survival were vascular invasion, age, ECOG score, AFP, albumin, creatinine, AST, extra-hepatic spread and aetiology. The model-predicted survival very similar to that observed. The Harrell's c-indices for training and validation sets were 0.72 and 0.70, respectively indicating good prediction. CONCLUSIONS: Our model ('PROSASH') predicts patient survival using baseline clinical features. However, it will require further validation in a routine clinical practice setting.

Development of pre and post-operative models to predict early recurrence of hepatocellular carcinoma after surgical resection.

BACKGROUND & AIMS: Resection is the most widely used potentially curative treatment for patients with early hepatocellular carcinoma (HCC). However, recurrence within 2 years occurs in 30-50% of patients, being the major cause of mortality. Herein, we describe 2 models, both based on widely available clinical data, which permit risk of early recurrence to be assessed before and after resection. METHODS: A total of 3,903 patients undergoing surgical resection with curative intent were recruited from 6 different centres. We built 2 models for early recurrence, 1 using preoperative and 1 using pre and post-operative data, which were internally validated in the Hong Kong cohort. The models were then externally validated in European, Chinese and US cohorts. We developed 2 online calculators to permit easy clinical application. RESULTS: Multivariable analysis identified male gender, large tumour size, multinodular tumour, high albumin-bilirubin (ALBI) grade and high serum alpha-fetoprotein as the key parameters related to early recurrence. Using these variables, a preoperative model (ERASL-pre) gave 3 risk strata for recurrence-free survival (RFS) in the entire cohort - low risk: 2-year RFS 64.8%, intermediate risk: 2-year RFS 42.5% and high risk: 2-year RFS 20.7%. Median survival in each stratum was similar between centres and the discrimination between the 3 strata was enhanced in the post-operative model (ERASL-post) which included 'microvascular invasion'. CONCLUSIONS: Statistical models that can predict the risk of early HCC recurrence after resection have been developed, extensively validated and shown to be applicable in the international setting. Such models will be valuable in guiding surveillance follow-up and in the design of post-resection adjuvant therapy trials. LAY SUMMARY: The most effective treatment of hepatocellular carcinoma is surgical removal of the tumour but there is often recurrence. In this large international study, we develop a statistical method that allows clinicians to estimate the risk of recurrence in an individual patient. This facility enhances communication with the patient about the likely success of the treatment and will help in designing clinical trials that aim to find drugs that decrease the risk of recurrence.

Risk estimation for biliary tract cancer: Development and validation of a prognostic score.

BACKGROUND & AIMS: Biliary tract cancer is a rare tumour entity characterized by a poor prognosis. We aimed to identify prognostic factors and create a prognostic score to estimate survival. METHODS: Clinical data of the training set, consisting of 569 patients treated from 2000 to 2010 at Hannover Medical School, were analysed. A prognostic model defining three prognostic risk groups was derived from Cox regression analyses. The score was applied and validated in an independent cohort of 557 patients from four different German centres. RESULTS: Median overall survival (OS) was 14.5 months. If complete resection was performed, the patients had a significantly improved OS (23.9 months; n=242) as compared to patients with non-resectable tumours (9.1 months; n=329, P<.0001). Based on univariable and multivariable analyses of clinical data, a prognostic model was created using variables available before treatment. Those were age, metastasis, C-reactive protein (CRP), international normalized ratio (INR) and bilirubin. The prognostic score distinguished three groups with a median OS of 21.8, 8.6 and 2.6 months respectively. The validation cohort had a median OS of 20.2, 14.0 and 6.5 months respectively. The prognostic impact of the score was independent of the tumour site and of treatment procedures. CONCLUSIONS: Here, we identified prognostic factors and propose a prognostic score to estimate survival, which can be applied to all patients independent of tumour site and before initial treatment. Further validation in prospective trials is required.

Impact of disease stage and aetiology on survival in hepatocellular carcinoma: implications for surveillance.

BACKGROUND: Variation in survival in hepatocellular carcinoma (HCC) has been attributed to different aetiologies or disease stages at presentation. While international guidelines recommend surveillance of high-risk groups to permit early diagnosis and curative treatment, the evidence that surveillance decreases disease-specific mortality is weak. METHODS: We compared HCC survival figures from Japan (n=1174) and Hong Kong (n=1675) over similar time periods (Japan 2000-2013, Hong Kong, China 2003-2014). The former has an intensive national surveillance programme, while the latter has none. We also analysed changes in survival in Japan over a 50-year period including data from before and after institution of a national HCC surveillance programme. RESULTS: In Japan, over 75% of cases are currently detected by surveillance, whereas in Hong Kong <20% of cases are detected presymptomatically. Median survival was 52 months in Japan and 17.8 months in Hong Kong; this survival advantage persisted after allowance for lead-time bias. Sixty-two per cent of Japanese patients had early disease at diagnosis and 63% received curative treatment. The comparable figures for Hong Kong were 31.7% and 44.1%, respectively. These differences could not be accounted for by disease aetiology, and patients in Hong Kong who were detected at an early stage had a similar survival to the analogous patients in Japan. CONCLUSIONS: The variation in survival is largely accounted for by stage at diagnosis, which in turn relates to the intensity of surveillance programmes and the consequent variation in curative therapeutic options.

Transarterial chemo-embolisation of hepatocellular carcinoma: impact of liver function and vascular invasion.

BACKGROUND: Transarterial chemo-embolisation (TACE) is recommended for patients with BCLC intermediate stage hepatocellular carcinoma (stage B), particularly in patients with good underlying liver function and minimal symptoms. The hepatoma arterial embolisation prognostic (HAP) score combines measures of liver function and tumour-related factors to offer a simple prognostic scoring system. The Albumin-Bilirubin (ALBI) grade permits assessment of the impact of liver function on survival. We aimed to investigate these two models and vascular invasion (VI). METHODS: In an international cohort of 3030 patients undergoing TACE, we examined the impact of liver function as assessed by the ALBI score, the HAP score and VI on survival. RESULTS: Classification according to ALBI grade resulted in non-overlapping survival curves in the overall data set and all regional cohorts. The HAP score was also validated. Tumour number, aetiology and VI were identified as additional independent prognostic risk factors not currently included in the HAP score. Survival was particularly poor for patients with VI. CONCLUSIONS: The ALBI grade categorised patients receiving TACE into three clear prognostic groups, thereby emphasising the importance of underlying liver function in the outcome of TACE. The HAP score has been validated internationally and the serious adverse impact of VI is clearly shown.

Impact of Viral Status on Survival in Patients Receiving Sorafenib for Advanced Hepatocellular Cancer: A Meta-Analysis of Randomized Phase III Trials.

Purpose Following the Sorafenib Hepatocellular Carcinoma Assessment Randomized Protocol (SHARP) trial, sorafenib has become the standard of care for patients with advanced unresectable hepatocellular carcinoma, but the relation between survival advantage and disease etiology remains unclear. To address this, we undertook an individual patient data meta-analysis of three large prospective randomized trials in which sorafenib was the control arm. Methods Of a total of 3,256 patients, 1,643 (50%) who received sorafenib were available. The primary end point was overall survival (OS). A Bayesian hierarchical approach for individual patient data meta-analyses was applied using a piecewise exponential model. Results are presented in terms of hazard ratios comparing sorafenib with alternative therapies according to hepatitis C virus (HCV) or hepatitis B virus (HBV) status. Results Hazard ratios show improved OS for sorafenib in patients who are both HBV negative and HCV positive (log [hazard ratio], -0.27; 95% CI, -0.46 to -0.06). Median unadjusted survival is 12.6 (11.15 to 13.8) months for sorafenib and 10.2 (8.88 to 12.2) months for "other" treatments in this subgroup. There was no evidence of improvement in OS for any other patient subgroups defined by HBV and HCV. Results were consistent across all trials with heterogeneity assessed using Cochran's Q statistic. Conclusion There is consistent evidence that the effect of sorafenib on OS is dependent on patients' hepatitis status. There is an improved OS for patients negative for HBV and positive for HCV when treated with sorafenib. There was no evidence of any improvement in OS attributable to sorafenib for patients positive for HBV and negative for HCV.