Artificial intelligence in medicine: mitigating risks and maximizing benefits via quality assurance, quality control, and acceptance testing

The adoption of artificial intelligence (AI) tools in medicine poses challenges to existing clinical workflows. This commentary discusses the necessity of context-specific quality assurance (QA), emphasizing the need for robust QA measures with quality control (QC) procedures that encompass (1) acceptance testing (AT) before clinical use, (2) continuous QC monitoring, and (3) adequate user training. The discussion also covers essential components of AT and QA, illustrated with real-world examples. We also highlight what we see as the shared responsibility of manufacturers or vendors, regulators, healthcare systems, medical physicists, and clinicians to enact appropriate testing and oversight to ensure a safe and equitable transformation of medicine through AI.

Male Runners With Lower Energy Availability Have Impaired Skeletal Integrity Compared to Nonathletes

CONTEXT

Female athletes, particularly runners, with insufficient caloric intake for their energy expenditure [low energy availability (EA) or relative energy deficiency] are at risk for impaired skeletal integrity. Data are lacking in male runners.

OBJECTIVE

To determine whether male runners at risk for energy deficit have impaired bone mineral density (BMD), microarchitecture, and estimated strength.

DESIGN

Cross-sectional.

SETTING

Clinical research center.

PARTICIPANTS

39 men (20 runners, 19 controls), ages 16-30 years.

MAIN OUTCOME MEASURES

Areal BMD (dual-energy x-ray absorptiometry); tibia and radius volumetric BMD and microarchitecture (high-resolution peripheral quantitative computed tomography); failure load (microfinite element analysis); serum testosterone, estradiol, leptin; energy availability.

RESULTS

Mean age (24.5 ± 3.8 y), lean mass, testosterone, and estradiol levels were similar; body mass index, percent fat mass, leptin, and lumbar spine BMD Z-score (-1.4 ± 0.8 vs -0.8 ± 0.8) lower (P < .05); and calcium intake and running mileage higher (P ≤ .01) in runners vs controls. Runners with EA

CONCLUSIONS

Despite weight-bearing activity, skeletal integrity is impaired in male runners with lower caloric intake relative to exercise energy expenditure, which may increase bone stress injury risk. Lower estradiol and lean mass are associated with lower tibial strength in runners.

Collapse

Key Words

• HR-pQCT
• athletes
• bone density
• energy availability

Collapse

MESH Headings

• Humans
• Male
• Female
• Leptin
• Cross-Sectional Studies
• Calcium
• Bone Density
• Absorptiometry, Photon
• Lumbar Vertebrae
• Testosterone
• Estradiol

Collapse

Grants

• K23 DK115903 NIDDK NIH HHS
• K24 HD071843 NICHD NIH HHS
• 1UL1TR002541-01 NIH HHS
• K24 HL092902 NHLBI NIH HHS
• Massachusetts General Hospital Executive Committee on Research
• National Institutes of Health
• Department of Defense Research Participation Program
• U.S. Army Research Institute of Environmental Medicine
• Oak Ridge Institute for Science and Education

Collapse

Affiliation(s)

Melanie S Haines Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Harvard Medical School, Boston, MA 02115, USA
Snimarjot Kaur Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Harvard Medical School, Boston, MA 02115, USA
Geetanjali Scarff Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA
Meghan Lauze Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA
Anu Gerweck Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA
Meghan Slattery Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA
Nicolas M Oreskovic Harvard Medical School, Boston, MA 02115, USA Department of Internal Medicine and Pediatrics, Massachusetts General Hospital, Boston, MA 02114, USA
Kathryn E Ackerman Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Harvard Medical School, Boston, MA 02115, USA Female Athlete Program, Division of Sports Medicine, Boston Children's Hospital, Boston, MA 02115, USA
Adam S Tenforde Harvard Medical School, Boston, MA 02115, USA Physical Medicine and Rehabilitation, Spaulding Rehabilitation Hospital, Cambridge, MA 02129, USA
Kristin L Popp Harvard Medical School, Boston, MA 02115, USA Endocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Military Performance Division, United States Army Research Institute of Environmental Medicine, Natick, MA 01760, USA Department of Energy, Oak Ridge Institute for Science and Education, Oak Ridge, TN 37830, USA
Mary L Bouxsein Harvard Medical School, Boston, MA 02115, USA Endocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Center for Advanced Orthopedic Studies, Beth Israel Deaconess Medical Center, Boston, MA 02115, USA
Karen K Miller Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Harvard Medical School, Boston, MA 02115, USA
Madhusmita Misra Neuroendocrine Unit, Massachusetts General Hospital, Boston, MA 02114, USA Harvard Medical School, Boston, MA 02115, USA Division of Pediatric Endocrinology, Massachusetts General Hospital, Boston, MA 02114, USA

Collapse

The Shigella Spp. Type III Effector Protein OspB Is a Cysteine Protease

The type III secretion system is required for virulence of many pathogenic bacteria. Bacterial effector proteins delivered into target host cells by this system modulate host signaling pathways and processes in a manner that promotes infection. Here, we define the activity of the effector protein OspB of the human pathogen Shigella spp., the etiological agent of shigellosis and bacillary dysentery. Using the yeast Saccharomyces cerevisiae as a model organism, we show that OspB sensitizes cells to inhibition of TORC1, the central regulator of growth and metabolism. In silico analyses reveal that OspB bears structural homology to bacterial cysteine proteases that target mammalian cell processes, and we define a conserved cysteine-histidine catalytic dyad required for OspB function. Using yeast genetic screens, we identify a crucial role for the arginine N-degron pathway in the yeast growth inhibition phenotype and show that inositol hexakisphosphate is an OspB cofactor. We find that a yeast substrate for OspB is the TORC1 component Tco89p, proteolytic cleavage of which generates a C-terminal fragment that is targeted for degradation via the arginine N-degron pathway; processing and degradation of Tco89p is required for the OspB phenotype. In all, we demonstrate that the Shigella T3SS effector OspB is a cysteine protease and decipher its interplay with eukaryotic cell processes. IMPORTANCEShigella spp. are important human pathogens and among the leading causes of diarrheal mortality worldwide, especially in children. Virulence depends on the Shigella type III secretion system (T3SS). Definition of the roles of the bacterial effector proteins secreted by the T3SS is key to understanding Shigella pathogenesis. The effector protein OspB contributes to a range of phenotypes during infection, yet the mechanism of action is unknown. Here, we show that S. flexneri OspB possesses cysteine protease activity in both yeast and mammalian cells, and that enzymatic activity of OspB depends on a conserved cysteine-histidine catalytic dyad. We determine how its protease activity sensitizes cells to TORC1 inhibition in yeast, finding that OspB cleaves a component of yeast TORC1, and that the degradation of the C-terminal cleavage product is responsible for OspB-mediated hypersensitivity to TORC1 inhibitors. Thus, OspB is a cysteine protease that depends on a conserved cysteine-histidine catalytic dyad.

Plasma ACE2 predicts outcome of COVID-19 in hospitalized patients

AIMS

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) binds to angiotensin converting enzyme 2 (ACE2) enabling entrance of the virus into cells and causing the infection termed coronavirus disease of 2019 (COVID-19). Here, we investigate associations between plasma ACE2 and outcome of COVID-19.

METHODS AND RESULTS

This analysis used data from a large longitudinal study of 306 COVID-19 positive patients and 78 COVID-19 negative patients (MGH Emergency Department COVID-19 Cohort). Comprehensive clinical data were collected on this cohort, including 28-day outcomes. The samples were run on the Olink® Explore 1536 platform which includes measurement of the ACE2 protein. High admission plasma ACE2 in COVID-19 patients was associated with increased maximal illness severity within 28 days with OR = 1.8, 95%-CI: 1.4-2.3 (P < 0.0001). Plasma ACE2 was significantly higher in COVID-19 patients with hypertension compared with patients without hypertension (P = 0.0045). Circulating ACE2 was also significantly higher in COVID-19 patients with pre-existing heart conditions and kidney disease compared with patients without these pre-existing conditions (P = 0.0363 and P = 0.0303, respectively).

CONCLUSION

This study suggests that measuring plasma ACE2 is potentially valuable in predicting COVID-19 outcomes. Further, ACE2 could be a link between COVID-19 illness severity and its established risk factors hypertension, pre-existing heart disease and pre-existing kidney disease.

Subclinical tuberculosis among adults with HIV: clinical features and outcomes in a South African cohort

BACKGROUND

Subclinical tuberculosis is an asymptomatic disease phase with important relevance to persons living with HIV. We describe the prevalence, clinical characteristics, and risk of mortality for HIV-infected adults with subclinical tuberculosis.

METHODS

Untreated adults with HIV presenting for outpatient care in Durban, South Africa were screened for tuberculosis-related symptoms and had sputum tested by acid-fast bacilli smear and tuberculosis culture. Active tuberculosis and subclinical tuberculosis were defined as having any tuberculosis symptom or no tuberculosis symptoms with culture-positive sputum. We evaluated the association between tuberculosis disease category and 12-month survival using Cox regression, adjusting for age, sex, and CD4 count.

RESULTS

Among 654 participants, 96 were diagnosed with active tuberculosis disease and 28 with subclinical disease. The median CD4 count was 68 (interquartile range 39-161) cells/mm3 in patients with active tuberculosis, 136 (72-312) cells/mm3 in patients with subclinical disease, and 249 (125-394) cells/mm3 in those without tuberculosis disease (P < 0.001). The proportion of smear positive cases did not differ significantly between the subclinical (29%) and active tuberculosis groups (14%, P 0.08). Risk of mortality was not increased in individuals with subclinical tuberculosis relative to no tuberculosis (adjusted hazard ratio 0.84, 95% confidence interval 0.26-2.73).

CONCLUSIONS

Nearly one-quarter of tuberculosis cases among HIV-infected adults were subclinical, which was characterized by an intermediate degree of immunosuppression. Although there was no significant difference in survival, anti-tuberculous treatment of subclinical cases was common.

TRIAL REGISTRATION

Prospectively registered on ClinicalTrials.gov , NCT01188941 (August 26, 2010).