Artificial intelligence predicts direct-acting antivirals failure among hepatitis C virus patients: A nationwide hepatitis C virus registry program.

BACKGROUND/AIMS: Despite the high efficacy of direct-acting antivirals (DAAs), approximately 1-3% of hepatitis C virus (HCV) patients fail to achieve a sustained virological response. We conducted a nationwide study to investigate risk factors associated with DAA treatment failure. Machine-learning algorithms have been applied to discriminate subjects who may fail to respond to DAA therapy. METHODS: We analyzed the Taiwan HCV Registry Program database to explore predictors of DAA failure in HCV patients. Fifty-five host and virological features were assessed using multivariate logistic regression, decision tree, random forest, eXtreme Gradient Boosting (XGBoost), and artificial neural network. The primary outcome was undetectable HCV RNA at 12 weeks after the end of treatment. RESULTS: The training (n=23,955) and validation (n=10,346) datasets had similar baseline demographics, with an overall DAA failure rate of 1.6% (n=538). Multivariate logistic regression analysis revealed that liver cirrhosis, hepatocellular carcinoma, poor DAA adherence, and higher hemoglobin A1c were significantly associated with virological failure. XGBoost outperformed the other algorithms and logistic regression models, with an area under the receiver operating characteristic curve of 1.000 in the training dataset and 0.803 in the validation dataset. The top five predictors of treatment failure were HCV RNA, body mass index, α-fetoprotein, platelets, and FIB-4 index. The accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of the XGBoost model (cutoff value=0.5) were 99.5%, 69.7%, 99.9%, 97.4%, and 99.5%, respectively, for the entire dataset. CONCLUSION: Machine learning algorithms effectively provide risk stratification for DAA failure and additional information on the factors associated with DAA failure.

Evaluation of the DNA stability of forensic markers used in betel-quid chewers' oral swab samples and oral cancerous specimens: implications for forensic application.

Chewed betel-quid (BQ) residues are often considered vital biological evidence at crime scenes, since the human DNA extracted from the residues is actually from buccal epithelial cells and can be associated with suspects. BQ-chewing is also a risk factor for oral diseases and/or cancers. Archived medical oral-specimens can be used to identify specific individuals under adverse conditions, although STR markers are known to be unstable in various tumor tissues. This study evaluates the DNA stability of forensic marker systems in BQ-chewers' oral epithelial cells, and in archived clinical specimens of oral cancer patients. The genotypes of oral and paired peripheral blood samples in 200 subjects were compared, using the commercialized typing systems of HLA-DQA1, PM (including LDLR, GYPA, HBGG, D7S8, and GC loci), and AmpFlSTR markers (including 9 STR loci and the Amelogenin gene). The 100 healthy BQ-chewers had consistent oral swab and paired blood sample genotypes analyzed withboth DQA1/PM and STR marker systems. In the 100 oral cancer patients, one discordant result at D7S8 was found in the 600DQA1/PM-marker loci, and 25 allelic alterations with expansion or contraction were detected in the 900 STR loci. The findings herein suggest that when cancerous specimens were tested, the HLA-DQA1/PM system with point polymorphism appears more reliable than the STR system with length polymorphism. Our results also indicate that healthy BQ-chewers' oral cotton swabs containing buccal epithelial cells are useful for forensic purposes using the HLA-DQA1, PM, and STR marker systems.

Allelic alterations at the STR markers in the buccal tissue cells of oral cancer patients and the oral epithelial cells of healthy betel quid-chewers: an evaluation of forensic applicability.

Although cancerous specimens are usually not used in forensic DNA typing, they might be forcibly employed under certain instances. On the other hand, though the oral epithelial samples have been applied to forensic identification, the great popularity of betel quid (BQ)-chewing in Taiwan, which is known to be a risk factor leading to an oral cancer, makes this application questionable. The DNA stability of nine short tandem repeat (STR) markers (the AmpFlSTR kit) was first investigated and then used to evaluate the forensic appropriateness of the oral samples of both healthy BQ-chewers and the archived clinical specimens from oral cancer patients. The analyses were performed on buccal samples from 100 BQ-chewers and 100 oral cancer patients, as well as their paired peripheral blood samples, and a group of 100 non-BQ-chewers were used for the control. In the group of 100 oral cancer patients, two types of DNA instability were found. They were major allelic imbalance, and allelic alterations including the expansion, the contraction and the un-classified type (i.e. can not be confirmed as the expansion or the contraction). The overall percentage of the cancerous subjects demonstrating DNA instability was 33% (five patients possessing both types of DNA instability). Both types of DNA instability showed a tendency of increasing with the severity of the pathological stage of oral cancer. Forty-four occurrences of major allelic imbalance were found from 21 cancer patients. The statistical result revealed that there was no significant difference in the allelic imbalanced occurrence among the nine STR loci. Allelic alterations were found in 17 patients, within which 12 individuals had the expansion, five had the contraction, and three were the un-classified type. Further, among these 17 patients, three were found to acquire multiple allelic alterations at multiple loci. In the group of 100 unrelated healthy BQ-chewers, two loci with major allelic imbalance were detected. However, the two imbalanced alleles were virtually half lost, and could still be recognized as heterozygous alleles. The statistical results of ANOVA, chi(2), and Scheffe tests indicated that the means of allelic imbalance at the nine STR loci of the oral cancerous group revealed a significant difference from those in the control group. Our results suggest that oral cancer tissues cannot be used as references for forensic purposes using the PCR-based STR systems, whereas the oral swabs from healthy BQ-chewers can be employed, but should be done with caution.