Automatic ganglion cell detection for improving the efficiency and accuracy of hirschprung disease diagnosis

Abstract

Histopathologic diagnosis of Hirschsprung's disease (HSCR) is time consuming and requires expertise. The use of artificial intelligence (AI) in digital pathology is actively researched and may improve the diagnosis of HSCR. The purpose of this research was to develop an algorithm capable of identifying ganglion cells in digital pathology slides and implement it as an assisting tool for the pathologist in the diagnosis of HSCR. Ninety five digital pathology slides were used for the construction and training of the algorithm. Fifty cases suspected for HSCR (727 slides) were used as a validation cohort. Image sets suspected to contain ganglion cells were chosen by the algorithm and then reviewed and scored by five pathologists, one HSCR expert and 4 non-experts. The algorithm was able to identify ganglion cells with 96% sensitivity and 99% specificity (in normal colon) as well as to correctly identify a case previously misdiagnosed as non-HSCR. The expert was able to achieve perfectly accurate diagnoses based solely on the images suggested by the algorithm, with over 95% time saved. Non-experts would require expert consultation in 20–58% of the cases to achieve similar results. The use of AI in the diagnosis of HSCR can greatly reduce the time and effort required for diagnosis and improve accuracy.

A new systematization of histological analysis for the diagnosis of Hirschsprung's disease

“…So, new studies have emerged aiming to simplify HD diagnosis, revisiting the use of H&E staining in rectal mucosa and submucosa fragments. 16 , 17 , 18 , 19 The detection of ganglion cells in the submucosal plexuses by utilizing H&E staining, as simple as the method might be, still creates a lot of controversies. Several authors cite the difficulty in identifying these cells, since they are diffusely distributed throughout the intestinal submucosa.…”

CircANKRD12/circTIMMDC1 act as miR-181b-5p sponge to synergistically regulate neural cell migration via the PROX1-NOTCH1-HES1 axis in Hirschsprung's disease

“…Hirschsprung's disease (HSCR) is the most common congenital alimentary tract developmental malformation in children with an incidence ranging between ~ 1/2000 and ~ 1/5000 live births [1,2]. The pathological characteristic of HSCR is completely lack of ganglion cells in submucosal and myenteric plexuses in colon [3], owing to the consequence of the premature arrest of the migration of ectodermal vagal neural crest cells to form the enteric nervous system (ENS) before 12th week of gestation [4,5]. This de ciency leads to peristaltic dysfunction and persistent tetanic contraction within the lesioned colon, nally causing intestinal obstruction [6].…”

Distribuição das células ganglionares em fragmentos de mucosa e submucosa retal para o diagnóstico da moléstia de Hirschsprung

“…Porém, novos trabalhos têm surgido com o objetivo de simplificar o diagnóstico revisitando o uso da coloração HE nestes fragmentos de mucosa e submucosa retal 8,35,48,50,60 .…”

“…In a recent study, we demonstrated that the analysis of 60 sections per fragment of rectal mucosa and submucosa stained by HE may ensure a 90% diagnostic accuracy. In this study, in 54% of the cases, neurons were found in the final sections (50)(51)(52)(53)(54)(55)(56)(57)(58)(59)(60). This encouraged us to study their distribution in the rectal submucosa, as a means to simplify the diagnosis.…”

“…Já através da análise da distribuição dos neurônios nos fragmentos com inervação normal pelo método estatístico de Bootstrapping, conseguimos, dentre os 60 cortes, escolher apenas 15 cortes. Verificamos que estes 15 cortes selecionados possuíam uma maior incidência de células ganglionares se comparados com os demais cortes, apresentando assim uma boa representação do material.Durante a aplicação deste modelo de sistematização nos 47 casos de suspeita de MH com a coloração HE, realizamos a leitura das lâminas observando apenas esses 15 cortes e obtivemos uma acurácia satisfatória de 93%.Na comparação dos resultados do HE com o nosso padrão ouro para o diagnóstico (AChE) houve apenas um caso de falso-positivo, e durante a reavaliação das lâminas verificamos que foi um erro de observador, pois havia uma célula ganglionar no corte 4, que foi um dos cortes selecionados para o modelo56,60 .Também houve dois casos em que foram visualizados neurônios na coloração HE, mas o diagnóstico pela técnica da AChE demostrava atividade da acetilcolinesterase. Estes dois casos foram revistos pela calretinina e foi confirmada a presença dos neurônios.Ao analisarmos o prontuário destes dois pacientes, verificamos que o paciente 4 possuía um laudo anatomopatológico da biopsia retal, realizado pela Divisão de Anatomia Patológica (DAP) do laboratório central do HCFMUSP, onde também não foram visualizadas células ganglionares.…”