The device, called the D3 (digital diffraction diagnosis) system, features an imaging module with a battery-powered LED light clipped onto a standard smartphone that records high-resolution imaging data with its camera.
With a much greater field of view than traditional microscopy, the D3 system is capable of recording data on more than 100,000 cells from a blood or tissue sample in a single image.
The data can then be transmitted for analysis to a remote graphic-processing server via a secure, encrypted cloud service and the results rapidly returned to the point of care.
Recommended: Smartphone health apps quite accurate: Study
"We believe the platform we have developed provides essential features at an extraordinary low cost," said Cesar Castro, co-lead author of the report.
For molecular analysis of tumours, a sample of blood or tissue is labelled with microbeads that bind to known cancer-related molecules and loaded into the D3 imaging module.
After the image is recorded and data transmitted to the server, the presence of specific molecules is detected by analysing the diffraction patterns generated by the microbeads.
A pilot test of the system with cancer cell lines detected the presence of tumour proteins with an accuracy matching that of the current gold standard for molecular profiling, and the larger field of view enabled simultaneous analysis of more than 100,000 cells at a time.
"D3 analysis promptly and reliably categorised biopsy samples as high-risk, low-risk or benign with results matching those of conventional pathologic analysis," the authors noted.
In the pilot tests, results of the D3 assay were available in under an hour and at a cost of $1.80 per assay, a price that would be expected to drop with further refinement of the system.
"We expect that the D3 platform will enhance the breadth and depth of cancer screening in a way that is feasible and sustainable for resource limited-settings," added Ralph Weissleder, director of the MGH Center for Systems Biology (CSB) and co-senior author of the paper.
The report appeared in the journal PNAS Early Edition.