Validation of MSIntuit as an AI-based pre-screening tool for MSI detection from colorectal cancer histology slides.

Mismatch Repair Deficiency (dMMR)/Microsatellite Instability (MSI) is a key biomarker in colorectal cancer (CRC). Universal screening of CRC patients for MSI status is now recommended, but contributes to increased workload for pathologists and delayed therapeutic decisions. Deep learning has the potential to ease dMMR/MSI testing and accelerate oncologist decision making in clinical practice, yet no comprehensive validation of a clinically approved tool has been conducted. We developed MSIntuit, a clinically approved artificial intelligence (AI) based pre-screening tool for MSI detection from haematoxylin-eosin (H&E) stained slides. After training on samples from The Cancer Genome Atlas (TCGA), a blind validation is performed on an independent dataset of 600 consecutive CRC patients. Inter-scanner reliability is studied by digitising each slide using two different scanners. MSIntuit yields a sensitivity of 0.96-0.98, a specificity of 0.47-0.46, and an excellent inter-scanner agreement (Cohen's κ: 0.82). By reaching high sensitivity comparable to gold standard methods while ruling out almost half of the non-MSI population, we show that MSIntuit can effectively serve as a pre-screening tool to alleviate MSI testing burden in clinical practice.

[Forensic autopsies and pathological examination: Current practice in Montpellier's jurisdiction. A retrospective study of 630 autopsies]. / Autopsies judiciaires et expertises anatomopathologiques complémentaires : état des pratiques sur le ressort de la cour d'appel de Montpellier. Étude rétrospective de 630 autopsies.

INTRODUCTION: In France, pathological examination is not systematically required in forensic autopsies. The factors affecting the decision to carry out a pathological expertise have not yet been defined. The aim of this study was to describe in which conditions a pathological expertise was required after forensic autopsy by the high court of Montpellier. METHODS: This study included and analyzed retrospectively all of the autopsy elements, of all forensic autopsies carried out over a year. These elements were classified: pre-autopsy, per-autopsy, and post-autopsy. RESULTS: A pathological expertise was required in 19.2% of 630 cases, among which 31% in a context of undetermined cause of death and in 14% of cases of determined causes of death. The forensic practitioner recommended a pathological expertise in 10 to 31% of autopsies. Overall, 64 pathological examinations were realized out of 121 recommended examinations (52.9%), this rate varied from 25 to 73% depending on the court. The magistrate tended to favor anatomopathological expertise in cases of determined causes of death, and in certain manner of death (80% homicide versus 35% natural). The pathologist's expertise enabled to change the cause of death in 22% of cases and the manner of death in 19%. The pathological approach was a major asset in the 65% of unknown manner of deaths and in the 20% of natural, whereas the expertise did not help in cases of homicides, suicides and accidents. The cause of death was modified in 5.6% of initially determined causes of death, against 42.9% in case of initially unknown cause. CONCLUSION: The use of pathologic examination in forensic autopsies is scarce and uneven. The factors resulting to its request are not directly linked to its scientific assets. A conjoint work between forensic and pathologist practitioners would be beneficial.

Cytokines as new biomarkers of skin wound vitality.

BACKGROUND: The diagnosis of skin wound vitality is currently based on standard histology, but histological findings lack sensitivity in case of a short survival time. New reliable biomarkers of vitality are therefore strongly needed. We assessed the ability of 10 candidate cytokines (IFN-γ, IL-1ß, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70, IL-13, TNF-α) to discriminate between vital and early post-mortem wounds. METHODS: Twenty-four cadavers with a recent open skin wound (< 3 h) were included (20 men, 4 women, mean age = 51.0 ± 24.3 years). An early post-mortem wound was performed in an uninjured skin area, and both wounds were sampled at the autopsy (post-mortem interval (PMI) = 66.3 ± 28.3 h). Needle-puncture sites related to resuscitation cares were included as very early post-mortem wounds (n = 6). In addition to standard histology, cytokines levels were simultaneously measured in each sample using a multiplex sandwich immunoassay, then normalized on healthy skin levels. A quantitative evaluation of IL-8-positive cells in ante- and post-mortem wound samples was also performed. RESULTS: In the training set of samples (n = 72), cytokine levels were significantly higher in vital wounds (mean age = 47 ± 53 min) than in post-mortem wounds (mean PMI = 6.9 ± 9.0 h) (p < 0.2), except for two cytokines (IFN-γ and IL-2). IL-8 was the best discriminatory cytokine (Se = 54%, Sp = 100%, AUC = 0.79), while a multivariate model combining IL-4 and IL12p70 was a bit more discriminant (Se = 55%, Sp = 100%, AUC = 0.84). In the validation set (n = 72), the discriminatory power of the cytokines and the predictive model was slightly lower, with IL-8 remaining the best cytokine (Se = 46%, Sp = 96%, AUC = 0.75). The predictive model remained highly specific (Sp = 100%). Both the cytokines and the predictive model allowed the iatrogenic injuries to be correctly classified as post-mortem wounds. Standard histology and immunohistochemistry showed 21% sensitivity and a specificity of 79% and 100%, respectively. Only two iatrogenic wounds could be properly categorized histologically. CONCLUSION: This study suggests that cytokines could be useful biomarkers of skin wound vitality and that the immunoassay method could be more sensitive than immunohistochemistry to identify wounds with a short survival time. Further research is underway to confirm these preliminary data.