Adrian Hood

Automatic normal-abnormal video frame classification for colonoscopy

Two novel schemes are proposed to represent intermediate-scale features for normal-abnormal classification of colonoscopy images. The first scheme works on the full-resolution image, the second on a multi-scale pyramid space. Both schemes support any feature descriptor; here we use multi-resolution local binary patterns which outperformed other features reported in the literature in our comparative experiments. We also compared experimentally two types of features not previously used in colonoscopy image classification, bag of features and sparse coding, each with and without spatial pyramid matching (SPM). We find that SPM improves performance, therefore supporting the importance of intermediate-scale features as in the proposed schemes for classification. Within normal-abnormal frame classification, we show that our representational schemes outperforms other features reported in the literature in leave-N-out tests with a database of 2100 colonoscopy images.

Quantitative assessment of colorectal morphology: Implications for robotic colonoscopy

This paper presents a method of characterizing the distribution of colorectal morphometrics. It uses three-dimensional region growing and topological thinning algorithms to determine and visualize the luminal volume and centreline of the colon, respectively. Total and segmental lengths, diameters, volumes, and tortuosity angles were then quantified. The effects of body orientations on these parameters were also examined. Variations in total length were predominately due to differences in the transverse colon and sigmoid segments, and did not significantly differ between body orientations. The diameter of the proximal colon was significantly larger than the distal colon, with the largest value at the ascending and cecum segments. The volume of the transverse colon was significantly the largest, while those of the descending colon and rectum were the smallest. The prone position showed a higher frequency of high angles and consequently found to be more torturous than the supine position. This study yielded a method for complete segmental measurements of healthy colorectal anatomy and its tortuosity. The transverse and sigmoid colons were the major determinant in tortuosity and morphometrics between body orientations. Quantitative understanding of these parameters may potentially help to facilitate colonoscopy techniques, accuracy of polyp spatial distribution detection, and design of novel endoscopic devices.

Use of a patient completed iPad questionnaire to improve pre-operative assessment

Developments in healthcare technology could improve patient care and reduce healthcare costs. There is a need to facilitate communication and increase efficiency in surgical pre-assessment clinics. This study aimed to develop an iPad application to deliver an electronic patient questionnaire, and to evaluate its use in the pre-assessment environment. Software was developed, MyOp, for a standard iPad that mirrored the paper-based pre-assessment system, with features designed for ease of patient use and remote data transfer. A case–control study was conducted, comparing use of MyOp with paper-based practice, to evaluate feasibility and patient preference. Patients were offered the use of MyOp or paper-based system. Outcomes measured included time to complete iPad questionnaire, consultation duration, and a patient preference questionnaire. MyOp cost £3500 to develop. 104 individuals participated in the study, 53 MyOp and 51 controls. MyOp reduced the median consultation duration by 5.00xa0min. A reduction was seen in all subgroups except those aged over 70 or urology patients. Patients preferred to complete the form independently, using a touchpad or computer but expressed concerns about data security. Use of an electronic patient questionnaire reduces consultation time delivering greater efficiency of pre-assessment nurse time. Preconceived ideas about the use of technology in older age groups are likely inaccurate and less of a barrier than previously thought. Electronic pre-assessments could be used routinely to reduce demands on healthcare facilities, improve patient care, and triage patients prior to clinic attendance.

Index admission laparoscopic cholecystectomy for patients with acute biliary symptoms: results from a specialist centre

BACKGROUNDnIndex admission laparoscopic cholecystectomy (ALC) is the treatment of choice for patients admitted with biliary symptoms but is performed in less than 15% of these admissions. We analysed our results for ALC within a tertiary hepatobiliary centre.nnnMETHODSnData from all cholecystectomies carried out under the care of the two senior authors from 1998 to 2008 were prospectively collected and interrogated.nnnRESULTSn1710 patients underwent cholecystectomy of which 439 (26%) were ALC. Patients operated on acutely did not have a significantly different complication rate (P= 0.279; 4% vs.3%). Factors predicting complications were abnormal alkaline phosphatase (ALP) (P= 0.037), dilated common bile duct (CBD) (P= 0.026), cholangitis (P= 0.040) and absence of on table cholangiography (OTC) (P= 0.011). There were no bile duct injuries. Patients undergoing ALC had a higher rate of conversion to an open procedure (P < 0.001:10% vs.3%). The proportion of complicated disease was higher in the ALC group (P < 0.001; 70% vs.31%). Only complicated disease (P= 0.006), absence of OTC (P < 0.001) and age greater than 65 years (P < 0.001) were predictive of conversion on multivariate analysis.nnnCONCLUSIONSnLaparoscopic cholecystectomy can be performed safely in patients with acute biliary symptoms and should be considered the gold standard for management of these patients thus avoiding avoidable readmissions and life-threatening complications. A higher conversion rate to an open procedure must be accepted when treating more complicated disease. It is the severity of disease rather than timing of surgery which most probably predicts complications and conversions.

Real-Time Measurement of the Tool-Tissue Interaction in Minimally Invasive Abdominal Surgery The First Step to Developing the Next Generation of Smart Laparoscopic Instruments

Introduction. Analysis of force application in laparoscopic surgery is critical to understanding the nature of the tool-tissue interaction. The aim of this study is to provide real-time data about manipulations to abdominal organs. Methods. An instrumented short fenestrated grasper was used in an in vivo porcine model, measuring force at the grasper handle. Grasping force and duration over 5 small bowel manipulation tasks were analyzed. Forces required to retract gallbladder, bladder, small bowel, large bowel, and rectum were measured over 30 seconds. Four parameters were calculated—T(hold), the grasp time; T(close), time taken for the jaws to close; F(max), maximum force reached; and F(rms), root mean square force (representing the average force across the grasp time). Results. Mean F(max) to manipulate the small bowel was 20.5 N (±7.2) and F(rms) was 13.7 N (±5.4). Mean T(close) was 0.52 seconds (±0.26) and T(hold) was 3.87 seconds (±1.5). In individual organs, mean F(max) was 49 N (±15) to manipulate the rectum and 59 N (±13.4) for the colon. The mean F(max) for bladder and gallbladder retraction was 28.8 N (±7.4) and 50.7 N (±3.8), respectively. All organs exhibited force relaxation, the F(rms) reduced to below 25 N for all organs except the small bowel, with a mean F(rms) of less than 10 N. Conclusion. This study has commenced the process of quantifying tool-tissue interaction. The static measurements discussed here should evolve to include dynamic measurements such as shear, torque, and retraction forces, and be correlated with evidence of histological damage to tissue.

Technological assessment of the biogalvanic method for tissue characterization

Biogalvanic cells have the potential to be used in characterizing biological tissue properties and ultimately tissue health. A biogalvanic cell is established by placing two differing metal electrodes across a target tissue allowing an electrical tissue-specific internal resistance to be determined. A novel data analysis method using least-squares fitting has been developed to more effectively determine the parameters of the biogalvanic system model proposed in the literature. The validity of the method has been examined through characterization of electrical models, ex vivo porcine tissue, and in vivo porcine tissue. Strong agreement between test results and the proposed characterization model has been shown. However, determined internal resistances are influenced by mechanical strain, current modulation direction and tissue thickness, indicating complexities at the electrode–tissue interface. These complexities undermine some assumptions upon which the biogalvanic model is based. Ultimately this technique could offer potential for use in minimally invasive surgery for discriminating tissue health but requires improved understanding and control of testing conditions.

An in vivo analysis of safe laparoscopic grasping thresholds for colorectal surgery

BackgroundAnalysis of safe laparoscopic grasping thresholds for the colon has not been performed. This study aimed to analyse tissue damage thresholds when the colon is grasped laparoscopically, correlating histological changes to mechanical compressive forces.MethodsAn instrumented laparoscopic grasper was used to measure the forces applied to porcine colon, with data captured and plotted as a force–time (f–t) curve. Haematoxylin and eosin histochemistry of tissue subjected to 10, 20, 40, 50 and 70xa0N for 5, 30 and 60xa0s was performed, and the area of colonic circular and longitudinal muscle was compared in grasped and un-grasped regions. The area under the f–t curve was calculated as a measure of the accumulated force applied, known as the force–time product (FTP).ResultsFTP ranged from 55.7 to 3793xa0N.s. Significant differences were observed between the muscle area of the grasped and un-grasped regions in both longitudinal and circular muscle at 50xa0N and above for all grasping times. For the longitudinal muscle, significant differences were observed between grasped and un-grasped areas at 20xa0N force for 30xa0s (mean differenceu2009=u200959 mm2, 95% CI 41–77 mm2, Pu2009=u20090.04), 20xa0N force for 60xa0s (mean differenceu2009=u200931 mm2, 95% CI 21.5–40.5 mm2, Pu2009=u20090.006) and 40xa0N force for 30xa0s (mean difference 37 mm2, 95% CI 27–47 mm2, Pu2009=u20090.006). Changes in histology correlated with mechanical forces applied to the longitudinal muscle at a FTP over 300xa0N s.ConclusionsThis study characterizes the grasping forces that result in histological changes to the colon and correlates these with a mechanical measurement of the applied force. The findings will contribute to the development of smart laparoscopic graspers with active constraints to prevent excessive grasping and tissue injury.