DeepTek employs Artificial Intelligence technology to enhance workflow for screening operations in outpatient or hospital ICU settings enabling quick screening of chest X-rays allowing experts to review such studies faster. For example, DeepTeks technology automatically screens digital X-rays for an enlarged heart, pleural effusion and pulmonary edema by analysing independently or concurrently from digital Chest X-Rays in under a minute.
Cadiomegaly and the role of AI in its objective assessments
Enlarged heart (Cardiomegaly) as it referred to in medical literature refers to the thickening of the heart muscles or the expansion of the chambers. Cardiomegaly is a sign of another heart issue, but not per se a disease. Coronary artery disease is the most common cause. Cardiomegaly can be caused by many conditions, including hypertension, coronary artery disease, infections, inherited disorders, and cardiomyopathies:
Enlarged heart as seen on chest radiographs can be quite subjective. The AI tool helps in objective assessment of the findings and it also quantifies the same and provides tools for easier analysis and faster reporting. The automated approach used to measure cardiothoracic ratio (CTR) makes reporting objective, structured and faster.
Unlike Cardiomegaly which can be ascertained from chest radiographs in the outpatient settings it can be silent and can even be detected incidentally in a regular health check-up or a public health screening. Pulmonary edema and pleural effusion could be seen in outpatient, inpatient and ICU settings.
Dr. Amit Kharat, Radiologist and Co-founder at DeepTek stated, “To address the two stages of the cardiomegaly detection challenge, we constructed different neural network models. The first model discriminated between X-rays taken from the AP and PA perspectives. To calculate the CTR, the second model used the heart and thoracic diameters from the PA X-rays.” Similarly for pleural fluid and pulmonary edema we trained the AI model on a substantially large set of annotated X-rays.”
Though these findings are diverse and there can be substantial overlap in findings on a chest radiograph the AI tool also assists in detecting 20+ other chest pathologies using such unique AI models. This recommendation can then be reviewed by experts in light of the clinical history and appropriate notes of the findings can now be documented in the reports by experts using smart reporting tools and smart tags. Ajit Patil, Co-founder of DeepTek stated, “We plan to equip our public health initiatives for TB screening with these additional AI tools so based on the requirement we can add value to population screening programs.”
Planning AI integration in the workflow can help experts receive notifications for the latter two conditions – pleural effusion and pulmonary edema or suggest a list of differentials. The AI tool integrated with smart radiology reports can add to these findings in radiology reports as well. The AI system takes care of the most basic manual tasks and automates the same, allowing experts to ensure faster reports, which are quantified, comprehensive and objective.
DeepTeks vision is to provide cutting-edge solutions based on deep learning algorithms to bridge the imaging sectors wide gap in terms of availability of radiologists at the right place and right time by using AI and its unique proposition of a smart AI first platform Augmento. Bridging this gap will empower imaging experts, radiologists, physicians, patients, governmental decision-makers, and non-profit organizations with the ability to systematize imaging workflow dynamics ensuring accessibility to medical imaging.
