Diagnostic Performance Measures Research Articles

A prevalence-robust measure of diagnostic test performance

A prevalence-robust measure of diagnostic test performance

The Diagnostic Value of Transthoracic Echocardiography for Pulmonary Hypertension: A Systematic Review and Meta-Analysis

Purpose Pulmonary hypertension (PH) is an important cause of morbidity and mortality. A commonly proposed non-invasive method in the diagnosis and assessment of PH is Transthoracic Echocardiography (TTE). In this systematic review and meta-analysis, our aim was to assess the accuracy of echocardiography in the diagnosis of PH. Methods The search was done up to the 31st of January 2023 using MEDLINE, Scopus, CENTRAL, and Web of Sciences databases. The studies were included if they were randomized controlled trials or observational in design and evaluated the diagnostic performance of TTE in diagnosing PH in comparison to right-sided heart catheterization as a reference method. The diagnostic performance measures included sensitivity, specificity, Diagnostic Odds Ratio (DOR), Positive Likelihood Ratio (PLR), and Negative Likelihood Ratio (NLR). Results The total number of the included patients was 4,523 from 38 articles. The overall sensitivity and specificity of echocardiography in diagnosing pulmonary hypertension were 54.8% (95%CI: 45.7%-63.6%) and 52.1% (95%CI: 39.0%-64.9%), respectively. Moreover, the pooled DOR was 1.321 (95%CI: 0.695-2.51). The pooled PLR and NLR were 1.145 (95%CI: 0.830-1.579) and 0.867 (95%CI: 0.628-1.197), respectively. The highest diagnostic performance of TTE was among group 1 and 3 PH patients. Conclusion Our analysis revealed that TTE had low diagnostic sensitivity, specificity, and accuracy. Due to the overall poor diagnostic performance of TTE, the diagnosis of PH and the assessment of response to therapies require right-sided heart catheterization. Future prospective studies to improve the diagnostic performance of TTE in the diagnosis of PH are needed.

Efficient modelling of infectious diseases in wildlife: A case study of bovine tuberculosis in wild badgers.

Bovine tuberculosis (bTB) has significant socio-economic and welfare impacts on the cattle industry in parts of the world. In the United Kingdom and Ireland, disease control is complicated by the presence of infection in wildlife, principally the European badger. Control strategies tend to be applied to whole populations, but better identification of key sources of transmission, whether individuals or groups, could help inform more efficient approaches. Mechanistic transmission models can be used to better understand key epidemiological drivers of disease spread and identify high-risk individuals and groups if they can be adequately fitted to observed data. However, this is a significant challenge, especially within wildlife populations, because monitoring relies on imperfect diagnostic test information, and even under systematic surveillance efforts (such as capture-mark-recapture sampling) epidemiological events are only partially observed. To this end we develop a stochastic compartmental model of bTB transmission, and fit this to individual-level data from a unique > 40-year longitudinal study of 2,391 badgers using a recently developed individual forward filtering backward sampling algorithm. Modelling challenges are further compounded by spatio-temporal meta-population structures and age-dependent mortality. We develop a novel estimator for the individual effective reproduction number that provides quantitative evidence for the presence of superspreader badgers, despite the population-level effective reproduction number being less than one. We also infer measures of the hidden burden of infection in the host population through time; the relative likelihoods of competing routes of transmission; effective and realised infectious periods; and longitudinal measures of diagnostic test performance. This modelling framework provides an efficient and generalisable way to fit state-space models to individual-level data in wildlife populations, which allows identification of high-risk individuals and exploration of important epidemiological questions about bTB and other wildlife diseases.

Impact of discordance in cardiac troponin T versus I on the performance of triage pathways for myocardial infarction

Abstract Background Biological, analytical, and regulatory differences between high-sensitivity cardiac troponin T (hs-cTnT) and hs-cTnI could lead to an underestimation of the performance of a hs-cTnT-based triage-algorithm if evaluated against a hs-cTnI-based reference standard, and vice versa. Unfortunately, the impact of hs-cTnT/hs-cTnI-discordance on diagnostic performance measures is largely unknown. Methods We quantified the impact of hs-cTnT/hs-cTnI-discordance on the diagnostic performance of the three most widely used and guideline recommended hs-cTnT/I-based triage algorithms for patients presenting with symptoms suggestive of myocardial infarction (MI): the ESC 0/1h-algorithm, the ESC 0/2h-algorithm and the High-STEACS pathway. For the central adjudication of the final diagnosis (reference standard), two independent cardiologists had access to all available clinical information and applied the 4th universal definition of MI, one time with serial hs-cTnT-Elecsys measurements (gold standard T) and a second time with serial hs-cTnI-Architect measurements (gold standard I). The primary diagnostic endpoint for the rule-out of NSTEMI was sensitivity, while for rule-in of NSTEMI specificity. Results Among 7728 eligible patients, the final diagnosis was NSTEMI in 1463 patients (18.9%). Of these, NSTEMI was diagnosed with only one of the two adjudication methods (either gold standard T or I) in 167 (11.4%) patients. When the three hs-cTnT-based early-triage algorithms were evaluated using gold standard T and gold standard I, sensitivity was very high and comparable. In contrast, when the three hs-cTnI-based early-triage algorithms were evaluated using gold standard T and gold standard I, sensitivity was again very high, but consistently lower when using gold standard T as compared to gold standard I. The difference in sensitivity was 1.1% (95%CI, 0.6–1.7) for ESC hs-cTnI-0/1h (Table 1A), 1.0% (95% CI, 0.5—1.6) for ESC hs-cTnI-0/2h (Table 1B), and 2.6% (95%CI, 1.7—3.5) for High-STEACS (Table 2). The differences observed in specificity mirrored those for sensitivity, i.e. always lower specificity when using the other hs-cTnT/I-assay for adjudication. Conclusion The impact of hs-cTnT/hs-cTnI-discordance on the diagnostic performance of hs-cTnT/I-based triage algorithms is modest, but large enough to result in a fallacious underestimation of the rule-out safety when evaluating hs-cTnI-based strategies using a hs-cTnT-based reference standard.

Anthropometric equation has sufficient diagnostic capacity to identify sarcopenia in women with breast cancer

BackgroundSarcopenia is a common condition in women with breast cancer, however still presents limitations for an effective diagnosis. This study aimed to evaluate the agreement and diagnostic accuracy of an anthropometric equation in diagnosing sarcopenia in women with breast cancer based on different constructs.MethodsCross-sectional study carried out with women with breast cancer aged ≥ 20 years. Sarcopenia was identified according to the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) criteria. Muscle strength was obtained by the handgrip strength test (HGS) and muscle mass (MM) by dual-energy x-ray absorptiometry (DXA) and by the anthropometric predictive equation. For the diagnosis of sarcopenia, eight constructs were proposed based on the MM assessment method (equation or DXA) and different cutoff points. Agreement analyses using Cohen’s Kappa test and diagnostic performance measures were performed. The significance level for all tests was 5%.ResultsA total of 122 women, with a mean age of 55.3 ± 11.4 years, were evaluated. There was a predominance of brown participants (50.8%), insufficiently active (57.4%), with diagnosis time ≤ 3 months (54.1%), and with invasive breast carcinoma (69.7%). The prevalence of sarcopenia ranged from 3.3 to 8.2%, depending on the construct used. The constructs determined from the cutoff points < 16.0 kg/< 7.58 m² and < 23.0 kg/< 7.58 m² were the ones that showed the best ability to detect sarcopenia.ConclusionThe anthropometric equation showed sufficient diagnostic capacity to be used as an alternative in identifying sarcopenia in women with breast cancer.

Artificial intelligence in commercial fracture detection products: a systematic review and meta-analysis of diagnostic test accuracy

Conventional radiography (CR) is primarily utilized for fracture diagnosis. Artificial intelligence (AI) for CR is a rapidly growing field aimed at enhancing efficiency and increasing diagnostic accuracy. However, the diagnostic performance of commercially available AI fracture detection solutions (CAAI-FDS) for CR in various anatomical regions, their synergy with human assessment, as well as the influence of industry funding on reported accuracy are unknown. Peer-reviewed diagnostic test accuracy (DTA) studies were identified through a systematic review on Pubmed and Embase. Diagnostic performance measures were extracted especially for different subgroups such as product, type of rater (stand-alone AI, human unaided, human aided), funding, and anatomical region. Pooled measures were obtained with a bivariate random effects model. The impact of rater was evaluated with comparative meta-analysis. Seventeen DTA studies of seven CAAI-FDS analyzing 38,978 x-rays with 8,150 fractures were included. Stand-alone AI studies (n = 15) evaluated five CAAI-FDS; four with good sensitivities (> 90%) and moderate specificities (80–90%) and one with very poor sensitivity (< 60%) and excellent specificity (> 95%). Pooled sensitivities were good to excellent, and specificities were moderate to good in all anatomical regions (n = 7) apart from ribs (n = 4; poor sensitivity / moderate specificity) and spine (n = 4; excellent sensitivity / poor specificity). Funded studies (n = 4) had higher sensitivity (+ 5%) and lower specificity (-4%) than non-funded studies (n = 11). Sensitivity did not differ significantly between stand-alone AI and human AI aided ratings (p = 0.316) but specificity was significantly higher the latter group (p < 0.001). Sensitivity was significant lower in human unaided compared to human AI aided respectively stand-alone AI ratings (both p ≤ 0.001); specificity was higher in human unaided ratings compared to stand-alone AI (p < 0.001) and showed no significant differences AI aided ratings (p = 0.316). The study demonstrates good diagnostic accuracy across most CAAI-FDS and anatomical regions, with the highest performance achieved when used in conjunction with human assessment. Diagnostic accuracy appears lower for spine and rib fractures. The impact of industry funding on reported performance is small.

Comparative Analysis of Somatosensory-Evoked Potentials and Transcranial Doppler Ultrasound for Cerebral Ischemia Detection in Carotid Endarterectomy: Insights from Network Meta-Analysis and Clinical Data

OBJECTIVEThis study aims to compare the diagnostic efficacy of somatosensory-evoked potentials (SEP) and transcranial Doppler sonography (TCD) for monitoring cerebral tissue ischemia during carotid endarterectomy (CEA) using network meta-analysis and retrospective analysis of clinical data. METHODSFor the meta-analysis, we conducted a comprehensive search of four electronic databases (PubMed, EMBASE, Cochrane, and Web of Science) from inception to September 2023, resulting in the inclusion of 52 relevant articles. Additionally, a retrospective study was conducted at our hospital, involving patients who underwent CEA surgery from July 2019 to July 2021. RESULTSThe network meta-analysis incorporated 52 articles, with ranking results indicating that SEP demonstrated superior performance in specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy , with surface under the cumulative ranking curve (SUCRA) values of 99.9%, 93.8%, 96.6%, and 99.9%, respectively. Furthermore, TCD exhibited the highest sensitivity with a SUCRA value of 92.0%. A total of 190 patients meeting inclusion criteria were included in the retrospective study. The area under the curve (AUC) for SEP's receiver operating characteristic (ROC) curve was 0.787, compared to TCD's AUC of 0.606. SEP demonstrated a sensitivity of 66.67%, with a specificity of 90.76%, PPV of 19.05%, NPV of 98.82%, and accuracy of 90%. For TCD, the diagnostic performance measures included a sensitivity of 50.00%, specificity of 71.19%, PPV of 5.35%, NPV of 97.76%, and accuracy of 70.53%. The Fisher's exact test for sensitivity yielded a result of P=1.000. The χˆ2 test for specificity resulted in χˆ2=22.863, with P<0.001. Continuous correction χˆ2 tests for PPV and NPV showed χˆ2=2.005 (P=0.157) and χˆ2=0.069 (P=0.793), respectively. Additionally, the χˆ2 test for accuracy showed χˆ2=22.742, with P<0.001. CONCLUSIONSDuring CEA, SEP appears to provide a slightly more reliable indication of the ischemic condition in cerebral tissues compared to TCD.

Peripheral facial palsy in children: Serum Borrelia antibodies may reduce the need for lumbar puncture.

We aimed to investigate the causes of acute peripheral facial palsy (PFP) in Danish children and to explore whether neuroborreliosis-related PFP could be diagnosed without lumbar puncture using clinical symptoms and serum Borrelia burgdorferi (Bb) antibodies. This retrospective population-based cohort study included children undergoing lumbar puncture for PFP between 2019 and 2023 in Denmark's Capital Region. Diagnostic performance measures for neuroborreliosis-related PFP were compared between serum Bb IgG alone and clinical risk scores combining Bb IgG with clinical parameters. Of the 326 patients with PFP, 137 (42%) were diagnosed with neuroborreliosis and 151 (46%) had Bell's palsy. Positive predictive value for serum Bb IgG alone was 88% (95% CI 79-93) and negative predictive value was 83% (95% CI 75-88). The positive predictive value of a risk score with seven additional parameters was 90% (95% CI 81-95) and negative predictive value 87% (95% CI 80-92). The positive predictive value of serum Bb IgG alone was high in our setting, where nearly half of children with PFP had neuroborreliosis. In high endemic settings, lumbar punctures may be reduced by (i) treating all children with PFP with doxycycline or (ii) treating Bb IgG positive children and performing lumbar puncture in seronegative children.

Prognostication Following Aneurysmal Subarachnoid Hemorrhage: The Modified Hunt and Hess Grading Scale

Background This study proposes a modification to the traditional Hunt and Hess (tHH) grading scale for prognostication in aneurysmal subarachnoid hemorrhage (aSAH), which differentiates the most severe‐grade patients based on the presence or absence of brainstem dysfunction [determined by Glasgow Coma Scale (GCS) scores 3‐5]. Methods Weighted aSAH hospitalizations were retrospectively identified in the National Inpatient Sample from 2015 to 2019 and were stratified by tHH and modified HH (mHH) grades. mHH grade 5 was defined as tHH grade 5 with GCS score 3–5, while mHH grade 4 comprised tHH grade 5 with GCS score 6–8 and tHH grade 4. HH grades 1–3 do not differ between the traditional and modified scales. Measures of diagnostic performance were compared for the primary study end point [poor outcome as determined by the previously validated NIS‐SAH Outcome Measure (NIS‐SOM), shown to have high concordance with modified Rankin Scale scores &gt; 2]. External validation of the mHH was performed using data from a prospectively maintained aSAH registry. Results Among 6130 aSAH hospitalizations, 2245 (36%) were tHH grade 5. Seven hundred and eighty‐five (35%) of these had a GCS 3–5 and were designated as mHH grade 5. Poor outcomes were identified in 78% and 77% of grade 4 tHH and mHH, respectively, and in 83% and 95% of grade 5 tHH and mHH, respectively. In comparison with the tHH, the mHH achieved superior discrimination [c‐statistic 0.793 (95% CI 0.768, 0.818) versus 0.780 (95% CI 0.750, 0.807); DeLong p &lt; 0.001] for poor outcome, as well as improved specificity (0.929 versus 0.304) and positive predictive value (PPV) (0.949 versus 0.827). External registry validation of the mHH demonstrated excellent discrimination [c‐statistic 0.835 (95% CI 0.801, 0.870)], with a specificity of 0.950 and PPV of 0.905. Conclusion The mHH achieved a favorable diagnostic performance profile using retrospective data and may aid in the prognostication of high‐severity patients with aSAH.

Ghost Sign on Diffusion-Weighted Imaging Generated Apparent Diffusion Coefficient Map: Additional MRI Diagnostic Marker for Extremity Osteomyelitis

Objective The aim of this study was to determine the sensitivity and specificity and inter-reader reliability of previously known “ghost sign” and “penumbra sign” on T1-weighted (T1W) imaging and “ghost sign” on apparent diffusion coefficient (ADC) map in osteomyelitis (OM) of the extremities. Materials and Methods In this cross-sectional retrospective study, two fellowship-trained musculoskeletal readers blinded to final diagnosis of OM versus no OM were asked to report the penumbra sign and ghost sign on T1W images and ghost sign on ADC map, as well as diagnosis of OM. Cohen's kappa was used. Diagnostic performance measures including sensitivity, specificity, and accuracy were calculated. Results A sample of 178 magnetic resonance imaging (MRI) scans of pathology-proven cases were included in this study, with 41 being positive for OM and 137 being negative for OM. There was a fair inter-reader agreement for imaging signs, and moderate agreement of 0.60 for OM. The sensitivities of the penumbra sign on T1W imaging, ghost sign on T1W imaging, and ghost sign on ADC map for OM are 3.7, 9.8, and 19.5%, respectively, while their respective specificities are 98.9, 97.8, and 94.5%, respectively. All three imaging signs showed a similar (good) accuracy of 76 to 78%. Conclusion The ghost sign on ADC can be used as an additional marker for OM and is a similarly highly specific but a more sensitive sign for OM than the conventionally used penumbra sign and ghost sign on T1W imaging. Key Points

Diagnostic Performance of 18F-FDG PET/CT According to Delay After Treatment to Detect Subclinical Recurrence of Head and Neck Squamous Cell Carcinoma.

Head and neck squamous cell carcinoma (HNSCC) remains a malignancy with high rates of locoregional recurrence and poor prognosis for recurrent cases. Early detection of subclinical lesions is challenging but critical for effective patient management. Imaging surveillance after treatment, particularly 18F-FDG PET/CT, has shown promise in the diagnosis of HNSCC recurrence. The aim was to evaluate the diagnostic performance of 18F-FDG PET/CT according to delay after treatment in detecting subclinical recurrence (SCR) in HNSCC patients. Methods: In this retrospective study, all 18F-FDG PET/CT scans were performed at a single center. All adults with histologically proven HNSCC who were treated with curative intent between January 1, 2006, and December 31, 2021, were included. They had a normal clinical examination before each scan. Patients who underwent an intensive follow-up strategy after treatment had 18F-FDG PET/CT with an intravenous contrast agent at 3-6 mo and annually thereafter for 5 y. The primary endpoint was diagnostic performance (positive and negative predictive values, sensitivity, specificity, and accuracy). Results: In total, 2,566 18F-FDG PET/CT scans were performed among 852 patients, with an average of 3 scans per patient. The overall diagnostic performance measures were as follows: positive predictive value (88%), negative predictive value (98%), sensitivity (98%), specificity (89%), and accuracy (93%). There were no significant differences in diagnostic performance over time. The scans detected 126 cases of SCR (14.8%) and 118 cases of metachronous cancer (13.8%). The incidence of SCR decreased over time, with the highest detection rate in the first 2 y after treatment. Positive predictive value improved over time, reaching 90% for the digital Vision 600 system (third period) compared with 76% for the analog Gemini GXLi system (first period, P < 0.001). Multivariate analysis identified advanced stage, high body mass index, and initial PET/CT upstaging as predictive factors for detection of SCR. Conclusion: Our study demonstrates that 18F-FDG PET/CT has high diagnostic performance in detecting SCR during follow-up after treatment of HNSCC, especially in the first 2 y. Advanced tumor stage, initial PET/CT upstaging, and high body mass index were associated with a higher likelihood of SCR detection. The routine use of 18F-FDG PET/CT during follow-up seems justified for patients with HNSCC.

Diagnostic accuracy of point-of-care cardiac ultrasound for acute coronary syndromes in patients presenting with chest pain to the emergency department: a single-center prospective study.

The diagnostic accuracy of focused cardiac ultrasound (FoCUS) performed in patients presenting to the emergency department (ED) with chest pain is currently unknown. The objective of this study was to assess the diagnostic accuracy of regional wall motion abnormalities detected with FoCUS for non-ST-elevation acute coronary syndrome (NSTE-ACS) diagnosis. A Single-center prospective observational study conducted in 2022 in the ED of the University Hospital Careggi, Italy. Adult patients presenting to the ED with acute nontraumatic chest pain were enrolled, irrespective of the presence of previous regional wall motion abnormalities. Patients with ST-segment elevation myocardial infarctions and patients with hemodynamic instability were excluded. FoCUS was performed at presentation by a trained ED physician. The final diagnosis of NSTE-ACS vs. alternative diagnosis was adjudicated by an ED physician blinded to FoCUS results after a 30-day follow-up. To assess if regional wall motion abnormalities were an independent predictor of NSTE-ACS, a multivariable logistic regression model was built. Diagnostic performance measures were calculated. A sensitivity analysis considering only type-1 NSTEMIs (i.e. plaque rupture/thrombosis) was conducted. Among 686 patients, NSTE-ACS was adjudicated in 106 (15.5%) patients, 67 of which were NSTEMIs. A total of 87 (12.7%) patients had regional wall motion abnormalities detected by FoCUS, which were an independent predictor of NSTE-ACS in the multivariable logistic regression analysis. Regional wall motion abnormalities had a sensitivity of 42.5% (33.0-51.9), a specificity of 92.8% (90.6-94.9), a negative predictive value of 89.8% (87.4-92.2), and a positive predictive value of 51.7% (41.2-62.2), for NSTE-ACS. Results were consistent in the sensitivity analysis. In ED patients with chest pain and no ST elevation, the detection of regional wall motion abnormalities was a predictor of NSTE-ACS. Despite a high specificity, which indicated a possible role of FoCUS in the rule-in of NSTE-ACS, sensitivity was too low to allow a safe rule-out using FoCUS results alone.

Influence of Covariates on 18F-FDG PET/CT Diagnostic Accuracy for Liver Metastasis.

(1) Background: 18F-FDG PET/CT diagnostic accuracy for liver metastasis (LM) could be influenced by technical parameters, lesion size, and the patient's covariates. This retrospective study aimed to evaluate these covariates' impact on PET/CT sensitivity. (2) Methods: Consecutive patients with suspected LMs who underwent 18F-FDG PET/CT scans were included. PET/CT scans were interpreted visually. The reference standard integrated histopathological and imaging follow-up. Logistic regression modeling and average marginal predictions were used to calculate per-lesion diagnostic performance measures with cluster robust 95% confidence intervals and to assess the covariates' impact on PET/CT sensitivity. (3) Results: We included 192 patients with 330 lesions. 18F-FDG PET/CT exhibited a per-lesion sensitivity, specificity, positive predictive value, and negative predictive value of 86%, 79%, 91%, and 69%, respectively. In multivariate analysis, TOF PET/CT exhibited a higher sensitivity than non-TOF PET/CT (91% vs. 78%, p = 0.02). Sensitivity was reduced for lesions < 10 mm compared to larger lesions (56% vs. 93%, p < 0.001). A 5 kg/m2 increase in BMI led to an average 5% sensitivity reduction (p < 0.001). Age, sex, blood glucose level below 11 mmol/L, and liver density did not significantly impact sensitivity (p > 0.05). (4) Conclusions: 18F-FDG PET/CT sensitivity might be reduced with non-TOF PET, lesions < 10 mm, and higher BMI.

Comparative Evaluation of Chest Ultrasonography and Computed Tomography as Predictors of Malignant Pleural Effusion: A Prospective Study.

Malignant pleural effusion (MPE) is a manifestation of advanced cancer that requires a prompt and accurate diagnosis. Ultrasonography (US) and computed tomography (CT) are valuable imaging techniques for evaluating pleural effusions; however, their relative predictive ability for a malignant origin remains debatable. This prospective study aimed to compare chest US with CT findings as predictors of malignancy in patients with undiagnosed exudative pleural effusion. Fifty-four adults with undiagnosed exudative pleural effusions underwent comprehensive clinical evaluation including chest US, CT, and histopathologic biopsy. Blinded radiologists evaluated the US and CT images for features suggestive of malignancy, based on predefined criteria. Diagnostic performance measures were calculated using histopathology as a reference standard. Of the 54 patients, 33 (61.1%) had MPEs confirmed on biopsy. No significant differences between US and CT were found in detecting parietal pleural abnormalities, lung lesions, chest wall invasion, or liver metastasis. US outperformed CT in identifying diaphragmatic pleural thickening ≥10 mm (33.3% vs. 6.1%, p < 0.001) and nodularity (45.5% vs. 3%, p < 0.001), whereas CT was superior for mediastinal thickening (48.5% vs. 15.2%, p = 0.002). For diagnosing MPE, diaphragmatic nodularity detected by US had 45.5% sensitivity and 100% specificity, whereas CT mediastinal thickening had 48.5% sensitivity and 90.5% specificity. Both US and CT demonstrate reasonable diagnostic performance for detecting MPE, with particular imaging findings favoring a malignant origin. US may be advantageous for evaluating diaphragmatic pleural involvement, whereas CT is more sensitive to mediastinal abnormalities.

White light computer-aided optical diagnosis of diminutive colorectal polyps in routine clinical practice.

Background and study aims Artificial Intelligence (AI) systems could make the optical diagnosis (OD) of diminutive colorectal polyps (DCPs) more reliable and objective. This study was aimed at prospectively evaluating feasibility and diagnostic performance of AI-standalone and AI-assisted OD of DCPs in a real-life setting by using a white light-based system (GI Genius, Medtronic Co, Minneapolis, Minnesota, United States). Patients and methods Consecutive colonoscopy outpatients with at least one DCP were evaluated by 11 endoscopists (5 experts and 6 non-experts in OD). DCPs were classified in real time by AI (AI-standalone OD) and by the endoscopist with the assistance of AI (AI-assisted OD), with histopathology as the reference standard. Results Of the 480 DCPs, AI provided the outcome "adenoma" or "non-adenoma" in 81.4% (95% confidence interval [CI]: 77.5-84.6). Sensitivity, specificity, positive and negative predictive value, and accuracy of AI-standalone OD were 97.0% (95% CI 94.0-98.6), 38.1% (95% CI 28.9-48.1), 80.1% (95% CI 75.2-84.2), 83.3% (95% CI 69.2-92.0), and 80.5% (95% CI 68.7-82.8%), respectively. Compared with AI-standalone, the specificity of AI-assisted OD was significantly higher (58.9%, 95% CI 49.7-67.5) and a trend toward an increase was observed for other diagnostic performance measures. Overall accuracy and negative predictive value of AI-assisted OD for experts and non-experts were 85.8% (95% CI 80.0-90.4) vs. 80.1% (95% CI 73.6-85.6) and 89.1% (95% CI 75.6-95.9) vs. 80.0% (95% CI 63.9-90.4), respectively. Conclusions Standalone AI is able to provide an OD of adenoma/non-adenoma in more than 80% of DCPs, with a high sensitivity but low specificity. The human-machine interaction improved diagnostic performance, especially when experts were involved.

Diagnosis of superficial endometriosis on transvaginal ultrasound by visualization of peritoneum of pouch of Douglas.

Around 80% of women with endometriosis have superficial endometriosis (SE) rather than ovarian or deep endometriosis (DE). However, to date, advances in non-invasive, imaging-based diagnosis have been limited to DE or ovarian disease. The objective of this study was to determine whether we can detect SE on transvaginal ultrasound scan (TVS) by assessing the peritoneum of the pouch of Douglas (POD). This was a retrospective diagnostic test study following a change in practice to include POD peritoneum assessment for SE during TVS at a tertiary London hospital. Eligible patients underwent TVS by a single clinician trained in endometriosis scanning and a subsequent surgical procedure (laparoscopy) between April 2018 and September 2021. Participants formed a consecutive series. The TVS findings were compared with those of laparoscopy as the gold standard. Comparison of TVS findings with intraoperative findings was performed by calculating the diagnostic test performance measures (sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and positive and negative likelihood ratios). The study included a total of 100 patients. We found that 43/100 (43.0%) patients had no endometriosis, 33/100 (33.0%) had SE and 24/100 (24.0%) had DE on laparoscopy. SE was correctly detected on TVS in 17/33 patients, with a sensitivity of 51.5% (95% CI, 33.5-69.2%), specificity of 94.0% (95% CI, 85.4-98.4%), PPV of 81.0% (95% CI, 60.8-92.1%) and NPV of 79.7% (95% CI, 73.4-84.9%). DE was correctly diagnosed in 20/24 cases, including all ovarian cases, with a sensitivity of 83.3% (95% CI, 62.3-95.3%), specificity of 97.4% (95% CI, 90.8-99.7%), PPV of 90.9% (95% CI, 71.6-97.5%) and NPV of 94.9% (95% CI, 88.3-97.8%). The detection of SE on TVS was most accurate in the POD (sensitivity, 50.0%; specificity, 96.4%; PPV, 76.9%; NPV, 88.9%). This study shows that the detection of SE in the POD is possible using routine TVS. While negative TVS does not reliably confirm the absence of disease or replace diagnostic laparoscopy, positive TVS facilitates non-invasive diagnosis for a much larger group of women than was previously possible. This should help to reduce the time from the onset of symptoms to diagnosis and enable initiation of medical treatment without the risk, cost and delay associated with a surgical diagnosis. © 2023 The Authors. Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.

Deep Learning and Gastric Cancer: Systematic Review of AI-Assisted Endoscopy.

Gastric cancer (GC), a significant health burden worldwide, is typically diagnosed in the advanced stages due to its non-specific symptoms and complex morphological features. Deep learning (DL) has shown potential for improving and standardizing early GC detection. This systematic review aims to evaluate the current status of DL in pre-malignant, early-stage, and gastric neoplasia analysis. A comprehensive literature search was conducted in PubMed/MEDLINE for original studies implementing DL algorithms for gastric neoplasia detection using endoscopic images. We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The focus was on studies providing quantitative diagnostic performance measures and those comparing AI performance with human endoscopists. Our review encompasses 42 studies that utilize a variety of DL techniques. The findings demonstrate the utility of DL in GC classification, detection, tumor invasion depth assessment, cancer margin delineation, lesion segmentation, and detection of early-stage and pre-malignant lesions. Notably, DL models frequently matched or outperformed human endoscopists in diagnostic accuracy. However, heterogeneity in DL algorithms, imaging techniques, and study designs precluded a definitive conclusion about the best algorithmic approach. The promise of artificial intelligence in improving and standardizing gastric neoplasia detection, diagnosis, and segmentation is significant. This review is limited by predominantly single-center studies and undisclosed datasets used in AI training, impacting generalizability and demographic representation. Further, retrospective algorithm training may not reflect actual clinical performance, and a lack of model details hinders replication efforts. More research is needed to substantiate these findings, including larger-scale multi-center studies, prospective clinical trials, and comprehensive technical reporting of DL algorithms and datasets, particularly regarding the heterogeneity in DL algorithms and study designs.

Implementing artificial intelligence-based cardio-obstetric screening in a diverse patient population: a prospective study

Abstract Background Cardiomyopathy is a leading cause of pregnancy-related mortality and represents the number one cause of death in the late postpartum period. Purpose The goal of this study was to prospectively evaluate the effectiveness of an artificial intelligence-based electrocardiogram (AI-ECG) and a point-of-care artificial intelligence-enhanced digital stethoscope for cardiomyopathy screening among pregnant and postpartum women. Methods We conducted a single-arm prospective study of pregnant and postpartum women (up to 12 months following delivery) enrolled at three sites between October 2021 and October 2022. Study participants completed a standard 12-lead ECG, ECG + phonocardiogram recordings with a digital stethoscope, a demographic questionnaire, and a transthoracic echocardiogram (TTE) within a 24-hour period. Diagnostic performance of the AI-ECG and digital stethoscope for cardiomyopathy detection, defined as left ventricular ejection fraction (LVEF) &amp;lt; 45%, was evaluated using the area under the curve (AUC) and standard measures of diagnostic performance. Results We included 100 consecutive pregnant and postpartum women who provided informed consent and had at least one ECG recorded, completed the questionnaire, and had an echocardiogram performed. All participants self-identified as female, and 41% were recruited from a federally qualified health center. The median age of participants was 31 years (Q1: 27, Q3: 34). Thirty-eight percent identified as non-Hispanic White, 32% as non-Hispanic Black, 21% as Hispanic or Latino, 6% as Asian, and 3% as Other (1 Native Hawaiian and 2 multiracial). Five percent (5/100) had LVEF &amp;lt;45%. The 12-lead AI-ECG model demonstrated close to perfect discrimination (accuracy 99%; AUC 1.00, sensitivity 100%) for detection of cardiomyopathy with 1 false positive screen (LVEF 49%), Figure 1A. The digital stethoscope (angled position) had an AUC of 0.99 (95% CI: 0.97, 1.00) for detection of LVEF &amp;lt;45%. Sensitivity, specificity, positive predictive value, and negative predictive value for the digital stethoscope were 60%, 98%, 60% and 98%, respectively for detection of LVEF &amp;lt;45% (Figure 1B &amp; 1C). Conclusions In this small prospective study, we found that 5% of pregnant and postpartum women had cardiomyopathy, a potentially life-threatening condition, and that a rapid, inexpensive, ubiquitous, massively scalable point-of-care test - the ECG - showed promising results, identifying it with an AUC &amp;gt; 0.98 with the application of artificial intelligence.Figure 1

Abstract 17209: Late Gadolinium Enhancement by Cardiovascular Magnetic Resonance versus Invasive Coronary Angiography: A Meta-Analysis on Diagnostic Accuracy for Cardiovascular Ischemia

Introduction: Invasive Coronary Angiography (ICA) has been the gold standard for diagnosing cardiovascular ischemia, however, it carries an inherent risk due to its invasive nature. Late Gadolinium Enhancement (LGE) by Cardiovascular Magnetic Resonance (CMR) is a promising non-invasive alternative. This necessitates determining its diagnostic accuracy to inform clinical decision-making. Hypothesis: This study aims to assess the diagnostic accuracy of LGE by CMR and ICA in detecting cardiovascular ischemia. Methods: We performed a systematic review of studies until January 2023 across multiple databases, including PubMed, Embase, and the Cochrane Library, that compared the diagnostic accuracy of LGE by CMR and ICA for cardiovascular ischemia. Diagnostic performance measures, including sensitivity, specificity, false-positive rate estimates, diagnostic odds ratio, and positive and negative Likelihood Ratios (LR), were pooled using a random-effects model and presented with a 95% Confidence Interval (CI). R software (version 4.0.3) with the mada package was utilized for statistical analyses. Results: Our meta-analysis collected data from 8 studies encompassing 1,020 patients, of which 395 had confirmed ischemic imaging and 625 had confirmed non-ischemic heart imaging. The results revealed that LGE by CMR had high diagnostic accuracy for detecting cardiovascular ischemia. The sensitivity was 90.9% (95% CI: 80.2%-96.1%, I2=50.4%), the specificity was 89.3% (95% CI: 79.2%-94.8%, I2=90.5%), and the false-positive rate was 10.7% (95% CI: 5.2%-20.8%). The diagnostic odds ratio for LGE by CMR was 82.82 (95% CI: 25.2-272.5), and the positive and negative Likelihood Ratios were 8.46 (95% CI: 4.19-17.06) and 0.10 (95% CI: 0.05-0.23), respectively. Conclusions: This study demonstrates the high diagnostic accuracy of LGE by CMR for cardiovascular ischemia, as indicated by its high sensitivity, specificity, and substantial diagnostic odds ratio. These findings suggest that LGE by CMR could be a reliable non-invasive alternative to ICA, potentially improving clinical decision-making and reducing risks associated with invasive imaging. Further studies and policy reviews are needed to optimize the use of this promising diagnostic modality.

Diagnostic accuracy of tumor M2-pyruvate kinase (tM2-PK) as a non-invasive biomarker in colorectal cancer: A systematic review and meta-analysis

IntroductionThe tumor pyruvate kinase M2 isoform (tM2-PK) is a glycolytic enzyme isoform that is present on the surface of rapidly proliferating cancer cells. The objective of this investigation was to assess the efficacy of the tM2-PK measurement assay in detecting colorectal cancer (CRC) through the analysis of serum/plasma and stool samples obtained from patients. MethodsThe pooled diagnostic performance measures, including sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), the area under the curve (AUC), Q*index, and summary receiver-operating characteristic curve (SROC), were computed using the Meta-Disc V.1.4 and Comprehensive Meta-Analysis V.3.3 software. The statistical methods of I2 and chi-square were employed to assess the presence of heterogeneity. The estimation of publication bias was conducted through the implementation of Begg's rank correlation and Egger's regression asymmetry tests. ResultsA total of 28 studies were found, involving 2900 participants (1560 cases and 1340 controls). The diagnostic accuracy of tM2-PK was calculated in CRC based on the pooled sensitivity of 83.70% (95% CI: 82.0% – 85.30%), specificity of 74.0% (95% CI: 72.0% – 76.0%), PLR of 4.432 (95% CI: 3.33 – 5.60), NLR of 0.187 (95% CI: 0.144 – 0.243), DOR of 30.182 (95% CI: 19.761 – 46.10) as well as AUC at 91.6%, and Q*-index at 85.0%. Publication bias was seen based on Begg's (p = 0.0006) and Egger's (p = 0.00015) tests. ConclusionThe results demonstrate that tM2-PK exhibits promise as a fair marker for CTRC, with the potential to serve as a non-invasive biomarker.