An investigation into augmentation and preprocessing for optimising X-ray classification in limited datasets: a case study on necrotising enterocolitis.

PURPOSE: Obtaining large volumes of medical images, required for deep learning development, can be challenging in rare pathologies. Image augmentation and preprocessing offer viable solutions. This work explores the case of necrotising enterocolitis (NEC), a rare but life-threatening condition affecting premature neonates, with challenging radiological diagnosis. We investigate data augmentation and preprocessing techniques and propose two optimised pipelines for developing reliable computer-aided diagnosis models on a limited NEC dataset. METHODS: We present a NEC dataset of 1090 Abdominal X-rays (AXRs) from 364 patients and investigate the effect of geometric augmentations, colour scheme augmentations and their combination for NEC classification based on the ResNet-50 backbone. We introduce two pipelines based on colour contrast and edge enhancement, to increase the visibility of subtle, difficult-to-identify, critical NEC findings on AXRs and achieve robust accuracy in a challenging three-class NEC classification task. RESULTS: Our results show that geometric augmentations improve performance, with Translation achieving +6.2%, while Flipping and Occlusion decrease performance. Colour augmentations, like Equalisation, yield modest improvements. The proposed Pr-1 and Pr-2 pipelines enhance model accuracy by +2.4% and +1.7%, respectively. Combining Pr-1/Pr-2 with geometric augmentation, we achieve a maximum performance increase of 7.1%, achieving robust NEC classification. CONCLUSION: Based on an extensive validation of preprocessing and augmentation techniques, our work showcases the previously unreported potential of image preprocessing in AXR classification tasks with limited datasets. Our findings can be extended to other medical tasks for designing reliable classifier models with limited X-ray datasets. Ultimately, we also provide a benchmark for automated NEC detection and classification from AXRs.

Laparoscopic inguinal hernia repair (LIHR): the benefit of the double stitch in the largest single-center experience.

AIM: To review our experience of laparoscopic inguinal hernia repair (LIHR) regarding complication rates, the practice of closing the asymptomatic patent processes vaginalis (PPV), and comparison of complication rates between pre-term (< 37 week gestation) and term infants. METHODS: Retrospective review of LIHR performed between 2009 and 2021. Repair was performed by intracorporal single or double purse string/purse string + Z-stitch using a non-absorbable suture. Data were analyzed using Chi-squared/Mann-Whitney and are quoted as median (range). RESULTS: 1855 inguinal rings were closed in 1195 patients (943 (79%) male). 1378 rings (74%) were symptomatic. 492 (41%) patients were pre-term. Corrected gestational age at surgery was 55 weeks (31 weeks-14.6 years) and weight 5.9 kg (1-65.5). Closure of contralateral PPV was higher in the premature group (210/397 [53%] vs. 265/613 [43%] p = 0.003). There were 23 recurrences in 20 patients, of whom 10 had been born prematurely. The only factor significantly associated with a lower recurrence was use of a second stitch (p = 0.011). CONCLUSION: This is the largest single-center reported series of LIHR. LIHR is safe at any age, the risk of recurrence is low, and can be corrected by re-laparoscopy. Use of a Z-stitch or second purse string is associated with a significantly lower rate of recurrence.

Maintaining a minimally invasive surgical service during a pandemic.

PURPOSE: The safety of minimally invasive surgery (MIS) was questioned in the COVID-19 pandemic due to concern regarding disease spread. We continued MIS during the pandemic with appropriate protective measures. This study aims to assess the safety of MIS compared to Open Surgery (OS) in this setting. METHODS: Operations performed during 2020 lockdown were compared with operations from the same time-period in 2019 and 2021. Outcomes reviewed included all complications, respiratory complications, length of stay (LOS) and operating surgeon COVID-19 infections (OSI). RESULTS: In 2020, MIS comprised 52% of procedures. 29% of MIS 2020 had complications (2019: 24%, 2021: 15%; p = 0.08) vs 47% in OS 2020 (p = 0.04 vs MIS). 8.5% of MIS 2020 had respiratory complications (2019: 7.7%, 2021: 6.9%; p = 0.9) vs 10.5% in OS 2020 (p = 0.8 vs MIS). Median LOS[IQR] for MIS 2020 was 2.5[6] days vs 5[23] days in OS 2020 (p = 0.06). In 2020, 2 patients (1.2%) were COVID-19 positive (MIS: 1, OS: 1) and there were no OSI. CONCLUSION: Despite extensive use of MIS during the pandemic, there was no associated increase in respiratory or other complications, and no OSI. Our study suggests that, with appropriate protective measures, MIS can be performed safely despite high levels of COVID-19 in the population.

Gastroschisis or exomphalos? An unusual abdominal defect.

A baby with an antenatal diagnosis of exomphalos was born at term. The abnormality had a highly unusual appearance, a right-sided paraumbilical defect, similar to gastroschisis, but with a sac typical of exomphalos containing both abdominal contents and over 1 L of serosanguinous fluid. The sac was drained and suspended from the top of the incubator in silo fashion. Definitive closure was achieved at day 3 of life, and the patient had a non-eventful recovery. Despite exomphalos and gastroschisis normally being discrete entities, ambiguity in the presentation this case necessitated a composite management approach.

C-reactive Protein is an Independent Predictor of Difficult Emergency Cholecystectomy.

Purpose The objective of this study was to identify variables that predict a difficult laparoscopic cholecystectomy performed in an emergency setting. The secondary aim was to devise a pathway for patients admitted acutely that required a cholecystectomy. Methods Patients admitted to the Emergency General Surgery Department at Nottingham, the United Kingdom that had an emergency cholecystectomy performed during the one-year period from May 2016 to June 2017 were identified. Collected data included patient demographics, clinical presentation, biochemical analysis, radiological findings, subsequent interventions, surgical data, and clinical outcome. A difficult cholecystectomy was defined as operative time >60 minutes, conversion to an open procedure, or sub-total cholecystectomy performed.  Results A total of 149 patients were included. Cholecystitis was the most common diagnosis (n = 86, 57.7%), followed by acute pancreatitis (n = 36, 24.1%). Fifty-five (36.9%) patients had an elevated C-reactive protein (CRP) >100 mg/dL. One hundred and twenty-one (81.2%) patients who had an emergency cholecystectomy were defined as "difficult". The overall morbidity rate was 15.4% (n = 23), and there was no post-operative in-hospital mortality. Univariate analysis showed that age >60 years (p = 0.012), underlying diagnosis (p = 0.010), presence of heart rate >90 (p = 0.027), and an elevated pre-surgery CRP >100 (p < 0.001) was associated with a difficult emergency cholecystectomy. Multi-variate analysis demonstrated that an elevated pre-surgery CRP >100 was an independent predictor of a difficult emergency cholecystectomy (p = 0.041). Conclusions An elevated pre-operative CRP is an independent predictor of a technically more difficult cholecystectomy in the emergency setting.