Reliability of Machine Learning in Eliminating Data Redundancy of Radiomics and Reflecting Pathophysiology in COVID-19 Pneumonia: Impact of CT Reconstruction Kernels on Accuracy

Background: Radiomical data are redundant but they might serve as a tool for lung quantitative assessment reflecting disease severity and actual physiological status of COVID-19 patients. Objective: Test the effectiveness of machine learning in eliminating data redundancy of radiomics and reflecting pathophysiologic changes in patients with COVID-19 pneumonia. Methods: We analyzed 605 cases admitted to Al Ain Hospital from 24 February to 1 July, 2020. They met the following inclusion criteria: age $\geq 18$ years;inpatient admission;PCR positive for SARS-CoV-2;lung CT available at PACS. We categorized cases into 4 classes: mild < 5% of pulmonary parenchymal involvement, moderate - 5-24%, severe - 25-49%, and critical $\geq50$ %. We used CT scans to build regression models predicting the oxygenation level, respiratory and cardiovascular functioning. Results: Radiomical findings are a reliable source of information to assess the functional status of patients with COVID-19. Machine learning models can predict the oxygenation level, respiratory and cardiovascular functioning from a set of demographics and radiomics data regardless of the settings of reconstructionkernels. The regression models can be used for scoring lung impairment and comparing disease severity in followup studies. The most accurate prediction we achieved was 6.454 +/- 3.715% of mean absolute error/range for all thefeatures and 7.069 +/- 4.17% for radiomics.Conclusion:The models may contribute to the proper risk evaluation anddisease management especially when the oxygen therapy impacts the actual values of the functional findings. Still,the structural assessment of an acute lung injury reflects the severity of the disease.

Reliability of machine learning in reflecting severity and pathophysiology in COVID-19 pneumonia: impact of CT reconstruction kernels on accuracy

Background: Radiomical data are redundant but they might serve as a tool for lung quantitative assessment reflecting disease severity and actual physiological status of COVID-19 patients. Objective: Test the effectiveness of machine learning in eliminating data redundancy of radiomics and reflecting pathophysiologic changes in patients with COVID-19 pneumonia. Methods: We analyzed 605 cases admitted to Al Ain Hospital from 24 February to 1 July, 2020. They met the following inclusion criteria: age≥18 years;inpatient admission;PCR positive for SARS-CoV-2;lung CT available at PACS. We categorized cases into 4 classes: mild ≤25% of pulmonary parenchymal involvement, moderate - 25-50%, severe - 50-75%, and critical –over 75%. We used CT scans to build regression models predicting the oxygenation level, respiratory and cardiovascular functioning. Results: Radiomical findings are a reliable source of information to assess the functional status of patients with COVID-19. Machine learning models can predict the oxygenation level, respiratory and cardiovascular functioning from a set of demographics and radiomics data regardless of the settings of reconstruction kernels. The regression models can be used for scoring lung impairment and comparing disease severity in follow up studies. The most accurate prediction we achieved was 6.454±3.715% of mean absolute error/range for all the features and 7.069±4.17% for radiomics. Conclusion: The models may contribute to the proper risk evaluation and disease management especially when the oxygen therapy impacts the actual values of the functional findings. Still, the structural assessment of an acute lung injury reflects the severity of the disease. Author

An anthropometric evaluation of the scapula, with emphasis on the coracoid process and glenoid fossa in a South African population.

The exact dimensions of the scapula, including the coracoid process and glenoid fossa, are fundamental in the patho-mechanics of the glenohumeral joint (GHJ); as these structures act as initiators of shoulder movement. The aim of the study was to evaluate the anthropometric parameters of the GHJ, with emphasis on the coracoid process and glenoid fossa. The morphometric (Linear Tools 2012, 0-150mm, LIN 86500963) and morphological parameters of a total of one hundred and sixty-four (n = 164) dry bone scapulae [Right (R): 80; Left (L): 84, Male (M): 68; Female (F): 96] were recorded. Results: (i) Shape of glenoid fossa: Type 1: (R) 16.5%, (L) 11.0%; Male (M) 20.1%, Female (F) 7.3%; Type 2: (R) 14.0%, (L) 15.2%; (M) 18.3%, (F) 11.0%; Type 3: (R) 18.3%, (L) 25.0%; (M) 27.4%, (F) 15.9%. (ii) Notch type: Type 1: (R) 1.7%, (L) 7.3%; (M) 6.7%, (F) 2.4%; Type 2: (R) 47.0%, (L) 43.9%; (M) 59.2%, (F) 31.7%. (iii) Vertical diameter of glenoid fossa (VD) (mm): (R) 35.2 ± 3.1, (L) 34.9 ± 3.0; (M) 35.3 ± 3.2, (F) 34.6 ± 2.8. (iv) Horizontal diameter 1 (HD1) of glenoid fossa (mm): (R) 18.4 ± 3.3, (L) 17.5 ± 2.9; (M) 18.2 ± 3.3, (F) 17.4 ± 2.6. (v) Horizontal diameter 2 (HD2) of glenoid fossa (mm): (R) 24.5 ± 2.9, (L) 23.6 ± 2.6; (M) 24.2 ± 2.7, (F) 23.7 ± 2.8. (vi) Length of coracoid process (CL) (mm): (R) 41.7 ± 4.7, (L) 41.5 ± 4.9; (M) 42.1 ± 4.7, (F) 40.7 ± 4.8. (vii) Width of coracoid process (CW) (mm): (R) 13.3 ± 1.9, (L) 14.2 ± 11.9; (M) 13.1 ± 1.9, (F) 15.1 ± 14.5. (viii) Coracoglenoid distance (CGD) (mm): (R) 27.4 ± 8.3, (L) 28.2 ± 3.5; (M) 28.2 ± 7.4, (F) 27.0 ± 3.4. In the present study, Type 3 (oval) was observed to be the predominant glenoid fossa shape with a higher incidence in male individuals and on the right side. Although only notch Types 1 (without a notch) and 2 (with one notch) were observed in this study, Type 2 (one notch) was the most prevalent, presenting with a significant p-value (p = 0.019), suggesting that notch Type 1 (without a notch) and 2 (with one notch) are common findings in the right and left side of individuals. The findings observed in this study may provide knowledge regarding the role of the coracoid parameters in etiology of subcoracoid impingement while knowledge on the glenoid fossa parameters and variations are essential for evaluation in shoulder arthroplasty for glenoid fractures and anterior dislocations, and for glenoid prosthesis designs for the South African population.

Interventions to improve headcount during covid-19 lockdown in South Africa

Background: The world was overcome by the COVID-19 pandemic from late 2019. South Africa entered into a country wide lockdown level 5 from March 26 to April 16, 2020. Public health facilities were greatly affected as they experienced reduced facility headcounts, which resulted in reduced HIV testing services (HTS), reduced patients attending their follow-up visits, and this also impacted the viral load completions. Methods: This was a retrospective review that analyzed the trends and the impact COVID-19 had on the headcount of primary health care (PHC) facilities and the number of patients accessing HTS and the Total Remaining on ART (TROA). In order to view the facility headcount and HTS trends on the same scale, for graphical representation the monthly figures have been indexed to their values in July 2019 (Jul 2019 = 100), prior to the impact of COVID-19 and the typical seasonal decline in activity during the holiday period. Results: Facility headcount dropped during the COVID-19 period (Mar 20, 0.98 to Apr 20, 0.73);and it is clear that the HTS trends (Mar 20, 0.89 to Apr 20, 0.47) mirror the headcount trends (Figure 1). However, the total remaining on ART remained relatively stable during this period;demonstrating successful programme efforts towards retention. These activities included case management of clients, community ART delivery, SMS reminders, extension of CCMDD (Centralised Chronic Medication Dispensing and Distribution) scripts, multi-month scripting and dispensing, improved appointment systems in facilities where we had filing interns, data quality improvement activities during this period and daily tracking with the district teams. Historically we have seen that HTS habitually drops during the holiday periods of December and April but starts picking up and follows the headcount trends. This year Level-5 lockdown brought on a steep drop with a strong recovery once lockdown restrictions eased, albeit not totally to former levels. Conclusion: Therefore, despite drastic drops in headcount from April 2020 to September 2020 as compared to 2019, the stability of TROA shows that implemented retention strategies have had a positive impact on the retention of patients during a pandemic.

Could COVID-19 Trigger a Rebirth in Anatomy Education? A Glimpse of Anatomists' Responses to Pandemics of the Past and Present.

Pandemics are nothing unusual but indeed lead to devastating effects that play a pivotal role in reshaping human history. The COVID-19 outbreak is currently responsible for major educational crises globally as most of the world has been faced with a mandated lockdown, and forced closure of educational institutions, including medical colleges. Anatomists have therefore been challenged to unlock technology in effort to achieve best deliverables for their discipline, without the use of traditional teaching aids such as the cadaver, osteological banks, prosected specimens, models, and microscopic slides. At present, the virtual classroom is the only option for the anatomist, thereby omitting vital aspects of the hidden curriculum such as ethical-reasoning, empathy, respect, professionalism, interpersonal, and communication skills. As body donations dwindle, the era of teaching in a cadaverless environment is upon us. This marks the beginning of a paradigm shift in education and research for anatomists worldwide. Given the variable pathological-morphological presentation in COVID-19-related deaths, it is also likely that the autopsy component of anatomic pathology will be resuscitated to demystify the underlying mechanisms of the virus. Since COVID-19 may never disappear completely, we would like to recommend that international anatomical societies collectively reach out to statutory bodies to devise a standardized method of teaching anatomy, employing readily available cost-effective resources, in the face of pandemics. However, if anatomy as a discipline has survived a millennium, surely anatomists can fight the "scourges" that have plagued them as various perspectives have been bandied about to welcome in a new normal.