Sickle cell disease (SCD) is a major public health priority throughout much of the world, affecting millions of people. In many regions, particularly those in resource-limited settings, SCD is not consistently diagnosed. In Africa, where the majority of SCD patients reside, more than 50% of the 0.2–0.3 million children born with SCD each year will die from it; many of these deaths are in fact preventable with correct diagnosis and treatment. Here, we present a deep learning framework which can perform automatic screening of sickle cells in blood smears using a smartphone microscope. This framework uses two distinct, complementary deep neural networks. The first neural network enhances and standardizes the blood smear images captured by the smartphone microscope, spatially and spectrally matching the image quality of a laboratory-grade benchtop microscope. The second network acts on the output of the first image enhancement neural network and is used to perform the semantic segmentation between healthy and sickle cells within a blood smear. These segmented images are then used to rapidly determine the SCD diagnosis per patient. We blindly tested this mobile sickle cell detection method using blood smears from 96 unique patients (including 32 SCD patients) that were imaged by our smartphone microscope, and achieved ~98% accuracy, with an area-under-the-curve of 0.998. With its high accuracy, this mobile and cost-effective method has the potential to be used as a screening tool for SCD and other blood cell disorders in resource-limited settings.

Peripheral blood smear (PBS) review by a pathologist is a necessary and invaluable diagnostic tool. However, innovative highly sophisticated haematology analysers that flag peripheral blood abnormalities have decreased the need for a PBS review. Ordering practices including PBS reviews lumped as part of an ‘order set’ or with complete blood count (CBC) constituted most PBS requests at our institution. A retrospective review of all PBS review orders from 1 April 2016 to 31 January 2017 was performed to investigate the ordering practices at our institution. A total of 2864 PBS were ordered during the above study period. In many cases, the PBS report did not add any significant clinical information beyond that acquired by the CBC and differential count. These findings inspired policy changes within our institution for pathologist PBS reviews. Within the electronic order system, all PBS orders for inpatients were linked to a pop-up window with criteria for peripheral smear review and instructions on the approval policy. Outpatient orders required clinicians to request pathology approval. This implementation reduced total number of PBS orders by 42.5% with no adverse effect on patient management. Empowering pathologists and clinicians with guidelines on PBS review orders is a beneficial educational exercise of resource utilisation. Discussion with physicians regarding clinical indications reduces non-contributory PBS reviews, provides guidance to appropriate testing, and aptly allocates pathologist and laboratory staff time and resources.

No abstract

We present a deep learning-based framework for performing automatic screening of sickle cells using a smartphone-based microscope. We achieved 98% accuracy when blindly testing 96 human blood smear slides, including 32 with sickle cell disease.

No abstract