Reporting guidelines in medical artificial intelligence: a systematic review and meta-analysis.

BACKGROUND: The field of Artificial Intelligence (AI) holds transformative potential in medicine. However, the lack of universal reporting guidelines poses challenges in ensuring the validity and reproducibility of published research studies in this field. METHODS: Based on a systematic review of academic publications and reporting standards demanded by both international consortia and regulatory stakeholders as well as leading journals in the fields of medicine and medical informatics, 26 reporting guidelines published between 2009 and 2023 were included in this analysis. Guidelines were stratified by breadth (general or specific to medical fields), underlying consensus quality, and target research phase (preclinical, translational, clinical) and subsequently analyzed regarding the overlap and variations in guideline items. RESULTS: AI reporting guidelines for medical research vary with respect to the quality of the underlying consensus process, breadth, and target research phase. Some guideline items such as reporting of study design and model performance recur across guidelines, whereas other items are specific to particular fields and research stages. CONCLUSIONS: Our analysis highlights the importance of reporting guidelines in clinical AI research and underscores the need for common standards that address the identified variations and gaps in current guidelines. Overall, this comprehensive overview could help researchers and public stakeholders reinforce quality standards for increased reliability, reproducibility, clinical validity, and public trust in AI research in healthcare. This could facilitate the safe, effective, and ethical translation of AI methods into clinical applications that will ultimately improve patient outcomes.

Artificial Intelligence (AI) refers to computer systems that can perform tasks that normally require human intelligence, like recognizing patterns or making decisions. AI has the potential to transform healthcare, but research on AI in medicine needs clear rules so caregivers and patients can trust it. This study reviews and compares 26 existing guidelines for reporting on AI in medicine. The key differences between these guidelines are their target areas (medicine in general or specific medical fields), the ways they were created, and the research stages they address. While some key items like describing the AI model recurred across guidelines, others were specific to the research area. The analysis shows gaps and variations in current guidelines. Overall, transparent reporting is important, so AI research is reliable, reproducible, trustworthy, and safe for patients. This systematic review of guidelines aims to increase the transparency of AI research, supporting an ethical and safe progression of AI from research into clinical practice.

Using histopathology latent diffusion models as privacy-preserving dataset augmenters improves downstream classification performance.

Latent diffusion models (LDMs) have emerged as a state-of-the-art image generation method, outperforming previous Generative Adversarial Networks (GANs) in terms of training stability and image quality. In computational pathology, generative models are valuable for data sharing and data augmentation. However, the impact of LDM-generated images on histopathology tasks compared to traditional GANs has not been systematically studied. We trained three LDMs and a styleGAN2 model on histology tiles from nine colorectal cancer (CRC) tissue classes. The LDMs include 1) a fine-tuned version of stable diffusion v1.4, 2) a Kullback-Leibler (KL)-autoencoder (KLF8-DM), and 3) a vector quantized (VQ)-autoencoder deploying LDM (VQF8-DM). We assessed image quality through expert ratings, dimensional reduction methods, distribution similarity measures, and their impact on training a multiclass tissue classifier. Additionally, we investigated image memorization in the KLF8-DM and styleGAN2 models. All models provided a high image quality, with the KLF8-DM achieving the best Frechet Inception Distance (FID) and expert rating scores for complex tissue classes. For simpler classes, the VQF8-DM and styleGAN2 models performed better. Image memorization was negligible for both styleGAN2 and KLF8-DM models. Classifiers trained on a mix of KLF8-DM generated and real images achieved a 4% improvement in overall classification accuracy, highlighting the usefulness of these images for dataset augmentation. Our systematic study of generative methods showed that KLF8-DM produces the highest quality images with negligible image memorization. The higher classifier performance in the generatively augmented dataset suggests that this augmentation technique can be employed to enhance histopathology classifiers for various tasks.

Direct image to subtype prediction for brain tumors using deep learning.

Background: Deep Learning (DL) can predict molecular alterations of solid tumors directly from routine histopathology slides. Since the 2021 update of the World Health Organization (WHO) diagnostic criteria, the classification of brain tumors integrates both histopathological and molecular information. We hypothesize that DL can predict molecular alterations as well as WHO subtyping of brain tumors from hematoxylin and eosin-stained histopathology slides. Methods: We used weakly supervised DL and applied it to three large cohorts of brain tumor samples, comprising N = 2845 patients. Results: We found that the key molecular alterations for subtyping, IDH and ATRX, as well as 1p19q codeletion, were predictable from histology with an area under the receiver operating characteristic curve (AUROC) of 0.95, 0.90, and 0.80 in the training cohort, respectively. These findings were upheld in external validation cohorts with AUROCs of 0.90, 0.79, and 0.87 for prediction of IDH, ATRX, and 1p19q codeletion, respectively. Conclusions: In the future, such DL-based implementations could ease diagnostic workflows, particularly for situations in which advanced molecular testing is not readily available.

The future landscape of large language models in medicine.

Large language models (LLMs) are artificial intelligence (AI) tools specifically trained to process and generate text. LLMs attracted substantial public attention after OpenAI's ChatGPT was made publicly available in November 2022. LLMs can often answer questions, summarize, paraphrase and translate text on a level that is nearly indistinguishable from human capabilities. The possibility to actively interact with models like ChatGPT makes LLMs attractive tools in various fields, including medicine. While these models have the potential to democratize medical knowledge and facilitate access to healthcare, they could equally distribute misinformation and exacerbate scientific misconduct due to a lack of accountability and transparency. In this article, we provide a systematic and comprehensive overview of the potentials and limitations of LLMs in clinical practice, medical research and medical education.

Generalizable biomarker prediction from cancer pathology slides with self-supervised deep learning: A retrospective multi-centric study.

Deep learning (DL) can predict microsatellite instability (MSI) from routine histopathology slides of colorectal cancer (CRC). However, it is unclear whether DL can also predict other biomarkers with high performance and whether DL predictions generalize to external patient populations. Here, we acquire CRC tissue samples from two large multi-centric studies. We systematically compare six different state-of-the-art DL architectures to predict biomarkers from pathology slides, including MSI and mutations in BRAF, KRAS, NRAS, and PIK3CA. Using a large external validation cohort to provide a realistic evaluation setting, we show that models using self-supervised, attention-based multiple-instance learning consistently outperform previous approaches while offering explainable visualizations of the indicative regions and morphologies. While the prediction of MSI and BRAF mutations reaches a clinical-grade performance, mutation prediction of PIK3CA, KRAS, and NRAS was clinically insufficient.

The effect of age, sex and BMI on the aldosterone-to-renin ratio in essential hypertensive individuals.

OBJECTIVE: The aldosterone-to-renin ratio (ARR) is widely used as a screening test for primary aldosteronism, but its determinants in patients with essential hypertension are not fully known. The purpose of the present investigation is to identify the impact of age, sex and BMI on renin, aldosterone and the ARR when measured under strict, standardized conditions in hypertensive patients without primary aldosteronism. METHODS: We analysed the data of 423 consecutive hypertensive patients with no concomitant cardiac or renal disorders from two different hospitals (Rotterdam and Maastricht) who had been referred for evaluation of their hypertension. Those who were diagnosed with secondary causes of hypertension, including primary aldosteronism, were excluded from analysis. Patients who used oral contraceptives or had hormonal replacement therapy were excluded as well. Plasma aldosterone concentration (PAC), active plasma renin concentration (APRC) and the ARR were measured under standardized conditions. All measurements were taken in the supine position at 10.00 h in the morning, with one subgroup of patients adhering to a sodium-restricted diet (55 mmol/day) for no less than 3 weeks, and the other subgroup maintaining an ad libitum diet. In those who were receiving antihypertensive treatment, all medications were discontinued at least 3 weeks before testing. RESULTS: In neither group did aldosterone correlate with age. Renin, however, was inversely related to age both during low-salt diet ( P  < 0.001) and during ad lib salt intake ( P  = 0.05). This resulted in a significant positive correlation between age and the ARR in both groups. Although on both dietary regimens, PAC and APRC were significantly higher in men when compared with women, the ARR was not significantly different between the two sexes. The age-relationships of renin and the ARR were comparable in men and women on both diets, albeit with greater variability in women. There was an upward trend between BMI and the ARR, which reached statistical significance only in men on low-salt diet. In multivariable regression analysis, age remained the only independent determinant of the ARR. CONCLUSION: In our essential hypertensive population, the ARR increased significantly with age but was not affected by sex or BMI.

The Effects of Verapamil, Hydralazine, and Doxazosin on Renin, Aldosterone, and the Ratio Thereof.

PURPOSE: Hydralazine, doxazosin, and verapamil are currently recommended by the Endocrine Society as acceptable bridging treatment in those in whom full cessation of antihypertensive medication is infeasible during screening for primary aldosteronism (PA). This is under the assumption that they cause minimal to no effect on the aldosterone-to-renin ratio, the most widely used screening test for PA. However, limited evidence is available regarding the effects of these particular drugs on said ratio. METHODS: In the present study, we retrospectively assessed the changes in aldosterone, renin, and aldosterone-to-renin values in essential hypertensive participants before and after treatment with either hydralazine (n = 26) or doxazosin (n = 20) or verapamil (n = 15). All samples were taken under highly standardized conditions. RESULTS: Hydralazine resulted in a borderline significant rise in active plasma renin concentration (19 vs 25 mIU/L, p = 0.067) and a significant fall in the aldosterone-to-renin ratio (38 vs 24, p = 0.017). Doxazosin caused declines in both plasma aldosterone concentration (470 vs 330 pmol/L, p = 0.028) and the aldosterone-to-renin ratio (30 vs 20, p = 0.020). With respect to verapamil, we found no statistically significant effect on any of these outcome variables. CONCLUSION: We conclude that the assumption that these drugs can be used with little consequence to the aldosterone-to-renin cannot be substantiated. While it is possible that they are indeed the best option when full antihypertensive drug cessation is infeasible, the potential effects of these drugs must still be taken into account when interpreting the aldosterone-to-renin ratio.

Variability of aldosterone, renin and the aldosterone-to-renin ratio in hypertensive patients without primary aldosteronism.

OBJECTIVES: The aldosterone-to-renin ratio (ARR) is commonly used in the screening of primary aldosteronism. However, limited information is available with regard to the intra-patient variability in this ratio. Our objective is to determine whether ARR measurements are reliably consistent over both the short- and long-term. METHODS: We assessed the short-term variability of the aldosterone-to-renin ratio in 116 unmedicated, essential hypertensive participants who had two blood samples taken in the morning of the same day for measurement of aldosterone and active plasma renin concentration. Long-term variability was studied in 22 unmedicated, essential hypertensive participants who had two blood samples taken approximately 1 year apart. All samples were taken under highly standardized conditions. RESULTS: Our data show that renin, aldosterone and the aldosterone-to-renin ratio show marked variations, both when measured on the same day and when assessed at a longer interval. The ARR becomes increasingly variable as its mean value increases. Its degree of variability is similar in both the short-term and the long-term. CONCLUSIONS: Based on our findings, we conclude that the aldosterone-to-renin has acceptable short-term variability in the lower ranges, but increasingly dubious reliability as aldosterone-to-renin values rise. Thus, in a clinical context, great caution should be taken in interpreting point-measurements of moderate to high aldosterone-to-renin ratio values.

The effect of medication on the aldosterone-to-renin ratio. A critical review of the literature.

The aldosterone-to-renin ratio (ARR) is a common screening test for primary aldosteronism in hypertensives. However, patients often use medications that could confound the ARR and, thereby, reduce the interpretability of the test. Since it is not always feasible to stop such medication, several drugs that are supposedly neutral with respect to the ARR have been recommended as alternative treatment. The objective of the present review is to explore whether sufficient evidence exists to justify the recommendations. To this end, we performed a systematic PubMed and Cochrane literature search regarding medications that may influence the ARR. Our review revealed that many commonly prescribed antihypertensives seem to have significant effects on renin, aldosterone, and resulting ARR values. However, the magnitude of these effects is poorly quantifiable with the present level of research. We conclude that several medications can affect the ARR. Not taking this into account could lead to misinterpretation of the ARR. Therefore, standardization of the medications used during ARR measurement is advisable for a reliable and accurate interpretation. Further research is needed to ascertain how to best optimize these medications.

Confounders of the aldosterone-to-renin ratio when used as a screening test in hypertensive patients: A critical analysis of the literature.

The aldosterone-to-renin ratio (ARR) is a common screening test for primary aldosteronism in hypertensives. However, there are many factors which could confound the ARR test result and reduce the accuracy of this test. The present review's objective is to identify these factors and to describe to what extent they affect the ARR. Our analysis revealed that sex, age, posture, and sodium-intake influence the ARR, whereas assay techniques do not. Race and body mass index have an uncertain effect on the ARR. We conclude that several factors can affect the ARR. Not taking these factors into account could lead to misinterpretation of the ARR.