Men diagnosed with azoospermia are told, in effect, that a laboratory looked at their semen sample and found no sperm. A newer approach argues that the problem may sometimes be the looking rather than the absence.
Systems combining high-speed microscopic imaging, machine learning, and microfluidics can scan millions of images from a single sample and flag cells that a human technician would not find in a realistic timeframe. A review of these methods from Columbia University Irving Medical Center appeared online in Current Opinion in Urology last month, summarizing the state of the field and concluding that further validation is required.
