Using saliva epigenetic data to develop and validate a multivariable predictor of esophageal cancer status.

Background: Salivary epigenetic biomarkers may detect esophageal cancer. Methods: A total of 256 saliva samples from esophageal adenocarcinoma patients and matched volunteers were analyzed with Illumina EPIC methylation arrays. Three datasets were created, using 64% for discovery, 16% for testing and 20% for validation. Modules of gene-based methylation probes were created using weighted gene coexpression network analysis. Module significance to disease and gene importance to module were determined and a random forest classifier generated using best-scoring gene-related epigenetic probes. A cost-sensitive wrapper algorithm maximized cancer diagnosis. Results: Using age, sex and seven probes, esophageal adenocarcinoma was detected with area under the curve of 0.72 in discovery, 0.73 in testing and 0.75 in validation datasets. Cancer sensitivity was 88% with specificity of 31%. Conclusion: We have demonstrated a potentially clinically viable classifier of esophageal cancer based on saliva methylation.

Development and validation of a multivariable risk factor questionnaire to detect oesophageal cancer in 2-week wait patients.

INTRODUCTION: Oesophageal cancer is associated with poor health outcomes. Upper GI (UGI) endoscopy is the gold standard for diagnosis but is associated with patient discomfort and low yield for cancer. We used a machine learning approach to create a model which predicted oesophageal cancer based on questionnaire responses. METHODS: We used data from 2 separate prospective cross-sectional studies: the Saliva to Predict rIsk of disease using Transcriptomics and epigenetics (SPIT) study and predicting RIsk of diSease using detailed Questionnaires (RISQ) study. We recruited patients from National Health Service (NHS) suspected cancer pathways as well as patients with known cancer. We identified patient characteristics and questionnaire responses which were most associated with the development of oesophageal cancer. Using the SPIT dataset, we trained seven different machine learning models, selecting the best area under the receiver operator curve (AUC) to create our final model. We further applied a cost function to maximise cancer detection. We then independently validated the model using the RISQ dataset. RESULTS: 807 patients were included in model training and testing, split in a 70:30 ratio. 294 patients were included in model validation. The best model during training was regularised logistic regression using 17 features (median AUC: 0.81, interquartile range (IQR): 0.69-0.85). For testing and validation datasets, the model achieved an AUC of 0.71 (95% CI: 0.61-0.81) and 0.92 (95% CI: 0.88-0.96) respectively. At a set cut off, our model achieved a sensitivity of 97.6% and specificity of 59.1%. We additionally piloted the model in 12 patients with gastric cancer; 9/12 (75%) of patients were correctly classified. CONCLUSIONS: We have developed and validated a risk stratification tool using a questionnaire approach. This could aid prioritising patients at high risk of having oesophageal cancer for endoscopy. Our tool could help address endoscopic backlogs caused by the COVID-19 pandemic.

Resistance to Antimalarial Monotherapy Is Cyclic.

Malaria is a prevalent parasitic disease that is estimated to kill between one and two million people-mostly children-every year. Here, we query PubMed for malaria drug resistance and plot the yearly citations of 14 common antimalarials. Remarkably, most antimalarial drugs display cyclic resistance patterns, rising and falling over four decades. The antimalarial drugs that exhibit cyclic resistance are quinine, chloroquine, mefloquine, amodiaquine, artesunate, artemether, sulfadoxine, doxycycline, halofantrine, piperaquine, pyrimethamine, atovaquone, artemisinin, and dihydroartemisinin. Exceptionally, the resistance of the two latter drugs can also correlate with a linear rise. Our predicted antimalarial drug resistance is consistent with clinical data reported by the Worldwide Antimalarial Resistance Network (WWARN) and validates our methodology. Notably, the cyclical resistance suggests that most antimalarial drugs are sustainable in the end. Furthermore, cyclic resistance is clinically relevant and discourages routine monotherapy, in particular, while resistance is on the rise. Finally, cyclic resistance encourages the combination of antimalarial drugs at distinct phases of resistance.

A critical evaluation of visual proportion of Gleason 4 and maximum cancer core length quantified by histopathologists.

Gleason score 7 prostate cancer with a higher proportion of pattern 4 (G4) has been linked to genomic heterogeneity and poorer patient outcome. The current assessment of G4 proportion uses estimation by a pathologist, with a higher proportion of G4 more likely to trigger additional imaging and treatment over active surveillance. This estimation method has been shown to have inter-observer variability. Fifteen patients with Prostate Grade Group (GG) 2 (Gleason 3 + 4) and fifteen patients with GG3 (Gleason 4 + 3) disease were selected from the PROMIS study with 192 haematoxylin and eosin-stained slides scanned. Two experienced uropathologists assessed the maximum cancer core length (MCCL) and G4 proportion using the current standard method (visual estimation) followed by detailed digital manual annotation of each G4 area and measurement of MCCL (planimetric estimation) using freely available software by the same two experts. We aimed to compare visual estimation of G4 and MCCL to a pathologist-driven digital measurement. We show that the visual and digital MCCL measurement differs up to 2 mm in 76.6% (23/30) with a high degree of agreement between the two measurements; Visual gave a median MCCL of 10 ± 2.70 mm (IQR 4, range 5-15 mm) compared to digital of 9.88 ± 3.09 mm (IQR 3.82, range 5.01-15.7 mm) (p = 0.64) The visual method for assessing G4 proportion over-estimates in all patients, compared to digital measurements [median 11.2% (IQR 38.75, range 4.7-17.9%) vs 30.4% (IQR 18.37, range 12.9-50.76%)]. The discordance was higher as the amount of G4 increased (Bias 18.71, CI 33.87-48.75, r 0.7, p < 0.0001). Further work on assessing actual G4 burden calibrated to clinical outcomes might lead to the use of differing G4 thresholds of significance if the visual estimation is used or by incorporating semi-automated methods for G4 burden measurement.

Development and validation of a risk prediction model to diagnose Barrett's oesophagus (MARK-BE): a case-control machine learning approach.

Background: Screening for Barrett's Oesophagus (BE) relies on endoscopy which is invasive and has a low yield. This study aimed to develop and externally validate a simple symptom and risk-factor questionnaire to screen for patients with BE. Methods: Questionnaires from 1299 patients in the BEST2 case-controlled study were analysed: 880 had BE including 40 with invasive oesophageal adenocarcinoma (OAC) and 419 were controls. This was randomly split into a training cohort of 776 patients and an internal validation cohort of 523 patients. External validation included 398 patients from the BOOST case-controlled study: 198 with BE (23 with OAC) and 200 controls. Identification of independently important diagnostic features was undertaken using machine learning techniques information gain (IG) and correlation based feature selection (CFS). Multiple classification tools were assessed to create a multi-variable risk prediction model. Internal validation was followed by external validation in the independent dataset. Findings: The BEST2 study included 40 features. Of these, 24 added IG but following CFS, only 8 demonstrated independent diagnostic value including age, gender, smoking, waist circumference, frequency of stomach pain, duration of heartburn and acid taste and taking of acid suppression medicines. Logistic regression offered the highest prediction quality with AUC (area under the receiver operator curve) of 0.87. In the internal validation set, AUC was 0.86. In the BOOST external validation set, AUC was 0.81. Interpretation: The diagnostic model offers valid predictions of diagnosis of BE in patients with symptomatic gastroesophageal reflux, assisting in identifying who should go forward to invasive testing. Overweight men who have been taking stomach medicines for a long time may merit particular consideration for further testing. The risk prediction tool is quick and simple to administer but will need further calibration and validation in a prospective study in primary care. Funding: Charles Wolfson Trust and Guts UK.

Machine Learning Creates a Simple Endoscopic Classification System that Improves Dysplasia Detection in Barrett's Oesophagus amongst Non-expert Endoscopists.

INTRODUCTION: Barrett's oesophagus (BE) is a precursor to oesophageal adenocarcinoma (OAC). Endoscopic surveillance is performed to detect dysplasia arising in BE as it is likely to be amenable to curative treatment. At present, there are no guidelines on who should perform surveillance endoscopy in BE. Machine learning (ML) is a branch of artificial intelligence (AI) that generates simple rules, known as decision trees (DTs). We hypothesised that a DT generated from recognised expert endoscopists could be used to improve dysplasia detection in non-expert endoscopists. To our knowledge, ML has never been applied in this manner. METHODS: Video recordings were collected from patients with non-dysplastic (ND-BE) and dysplastic Barrett's oesophagus (D-BE) undergoing high-definition endoscopy with i-Scan enhancement (PENTAX®). A strict protocol was used to record areas of interest after which a corresponding biopsy was taken to confirm the histological diagnosis. In a blinded manner, videos were shown to 3 experts who were asked to interpret them based on their mucosal and microvasculature patterns and presence of nodularity and ulceration as well as overall suspected diagnosis. Data generated were entered into the WEKA package to construct a DT for dysplasia prediction. Non-expert endoscopists (gastroenterology specialist registrars in training with variable experience and undergraduate medical students with no experience) were asked to score these same videos both before and after web-based training using the DT constructed from the expert opinion. Accuracy, sensitivity, and specificity values were calculated before and after training where p < 0.05 was statistically significant. RESULTS: Videos from 40 patients were collected including 12 both before and after acetic acid (ACA) application. Experts' average accuracy for dysplasia prediction was 88%. When experts' answers were entered into a DT, the resultant decision model had a 92% accuracy with a mean sensitivity and specificity of 97% and 88%, respectively. Addition of ACA did not improve dysplasia detection. Untrained medical students tended to have a high sensitivity but poor specificity as they "overcalled" normal areas. Gastroenterology trainees did the opposite with overall low sensitivity but high specificity. Detection improved significantly and accuracy rose in both groups after formal web-based training although it did it reach the accuracy generated by experts. For trainees, sensitivity rose significantly from 71% to 83% with minimal loss of specificity. Specificity rose sharply in students from 31% to 49% with no loss of sensitivity. CONCLUSION: ML is able to define rules learnt from expert opinion. These generate a simple algorithm to accurately predict dysplasia. Once taught to non-experts, the algorithm significantly improves their rate of dysplasia detection. This opens the door to standardised training and assessment of competence for those who perform endoscopy in BE. It may shorten the learning curve and might also be used to compare competence of trainees with recognised experts as part of their accreditation process.

"Stoneless" or Negative Ureteroscopy: A Reality in the Endourologic Routine or Avoidable Source of Frustration? Estimating the Risk Factors for a Negative Ureteroscopy.

INTRODUCTION: A "Negative" ureteroscopy (URS) is defined as a URS in which no stone is found during the procedure. It may occur when the stone has already been passed spontaneously or when it is located outside the collecting system. The aim of the study was to outline risk factors for Negative-URS. MATERIALS AND METHODS: We retrospectively analyzed the possible risk factors for Negative-URS from a database of 341 URS cases. In every case where presumptive ureteral stone was not found, a formal nephroscopy as well as a whole collecting system revision was completed. The Negative-URS group was compared with the non-Negative-URS group, in terms of patient and stone characteristics. RESULTS: The database of 341 URS cases included 448 different stone instances, of which 17 (3.8%) were negative and 431 (96.2%) were therapeutic. There was no statistical significant difference between the two groups concerning age, body mass index, stone location in the ureter, stone laterality, and whether the patient was prestented. The stepwise multiple logistic regression revealed three important risk factors, namely CT stone surface area (p < 0.0001), radiopacity of the stone at kidney, ureter, and bladder radiograph (KUB; p = 0.0004), and gender (p = 0.0011) with an area under the curve of 0.91. Women were found to have more possibilities to have a negative procedure by four- to sevenfold than men depending on the model. A nonradio-opaque stone at KUB is more likely to be correlated with a Negative-URS by 9.5- to 11-fold more than a radiopaque stone at KUB. For each increase of 1 U in CT stone surface area, there is an increase of 10%-12% to be non-negative. CONCLUSIONS: Female gender, a nonradio-opaque stone at KUB, and a smaller stone surface were statistically significantly different in the Negative-URS population.