LAHORE: Punjab Chief Minister Maryam Nawaz Sharif has taken an in-principle decision to introduce an “AI-Powered Multi-Disease Scanner” for patients in the province, and has directed Health Department officials to prepare a pilot project.
The decision followed a detailed briefing on the artificial-intelligence-equipped medical scanner given to the chief minister during a meeting with executives of the Chinese company Alibaba.
The briefing set out that the scanner uses contrast CT imaging to identify multiple diseases in approximately two minutes. Officials were told the technology assists across screening, diagnosis, treatment and disease management.
The system is designed to enable timely screening and identification of several forms of lethal cancer, including pancreatic, liver, gastric, oesophageal and colorectal cancers.
Beyond oncology, the scanner is intended to support diagnosis of Acute Aortic Syndrome, a group of life-threatening conditions affecting the main artery; cardiovascular disease risk assessment; and fat and muscle quantification, which measures the proportions of fat and muscle tissue in the body.
The briefing added that the technology also assists in osteoporosis assessment, the detection of reduced bone density, and fatty liver assessment.
Officials told the meeting that the AI-powered multi-disease scanner represents a significant advance in making medical diagnosis faster and more effective. Critically, the briefing stated that the system will function as an assistive technology alongside conventional diagnostic methods, rather than replacing them.
Maryam Nawaz Sharif expressed particular interest in the technology presented by the Alibaba executives and took a detailed briefing on the scanner before ordering the Health Department to prepare a pilot.
The clinical concept behind a multi-disease scanner is worth explaining, because it is not simply a faster CT machine.
Conventional CT imaging is ordered to answer one question. A patient sent for a scan after an abdominal complaint is assessed for that complaint; the enormous volume of additional anatomical information captured in the same images — bone density, liver fat content, vascular calcification, muscle mass — is generally not analysed, because doing so would require radiologist hours that no health system has spare.
AI-assisted multi-disease analysis targets exactly that unused data. The same scan, already acquired and already paid for, is passed through algorithms that flag findings across several organ systems at once. In health-economics terms this is called opportunistic screening, and its appeal in a resource-constrained system is obvious: it adds diagnostic yield without adding scans, radiation exposure or patient visits.
The caveats are equally well documented. Opportunistic screening produces incidental findings, and incidental findings generate follow-up investigations — some of which turn out to be unnecessary. A system that flags more requires the downstream capacity to resolve what it flags. That is a workforce and pathology-capacity question, not a software question.
Three things will determine whether the pilot succeeds. First, validation: algorithms trained predominantly on one population do not automatically perform equally on another, and the pilot should publish sensitivity and specificity figures for a Punjabi patient cohort rather than relying on vendor data. Second, the contrast-CT prerequisite: this technology requires functioning CT infrastructure, contrast media supply and trained radiographers at the site, which narrows the realistic pilot sites to tertiary hospitals. Third, data governance — a Chinese-vendor AI system processing patient imaging raises questions about where images are processed and stored, and under what legal framework, that Pakistan’s health regulations do not currently answer.
The chief minister’s own briefing struck the right note in describing the scanner as assistive rather than substitutive. Holding to that framing, and publishing the pilot’s results, would make this a genuinely serious project rather than a procurement announcement.
Leave a Reply