Multicontrast Pocket Colposcopy Cervical Cancer Diagnostic Algorithm for Referral Populations.

Objective and Impact Statement. We use deep learning models to classify cervix images-collected with a low-cost, portable Pocket colposcope-with biopsy-confirmed high-grade precancer and cancer. We boost classification performance on a screened-positive population by using a class-balanced loss and incorporating green-light colposcopy image pairs, which come at no additional cost to the provider. Introduction. Because the majority of the 300,000 annual deaths due to cervical cancer occur in countries with low- or middle-Human Development Indices, an automated classification algorithm could overcome limitations caused by the low prevalence of trained professionals and diagnostic variability in provider visual interpretations. Methods. Our dataset consists of cervical images (n=1,760) from 880 patient visits. After optimizing the network architecture and incorporating a weighted loss function, we explore two methods of incorporating green light image pairs into the network to boost the classification performance and sensitivity of our model on a test set. Results. We achieve an area under the receiver-operator characteristic curve, sensitivity, and specificity of 0.87, 75%, and 88%, respectively. The addition of the class-balanced loss and green light cervical contrast to a Resnet-18 backbone results in a 2.5 times improvement in sensitivity. Conclusion. Our methodology, which has already been tested on a prescreened population, can boost classification performance and, in the future, be coupled with Pap smear or HPV triaging, thereby broadening access to early detection of precursor lesions before they advance to cancer.

A novel speculum-free imaging strategy for visualization of the internal female lower reproductive system.

Fear of the speculum and feelings of vulnerability during the gynecologic exams are two of the biggest barriers to cervical cancer screening for women. To address these barriers, we have developed a novel, low-cost tool called the Callascope to reimagine the gynecological exam, enabling clinician and self-imaging of the cervix without the need for a speculum. The Callascope contains a 2 megapixel camera and contrast agent spray mechanism housed within a form factor designed to eliminate the need for a speculum during contrast agent administration and image capture. Preliminary bench testing for comparison of the Callascope camera to a $20,000 high-end colposcope demonstrated that the Callascope camera meets visual requirements for cervical imaging. Bench testing of the spray mechanism demonstrates that the contrast agent delivery enables satisfactory administration and cervix coverage. Clinical studies performed at Duke University Medical Center, Durham, USA and in Greater Accra Regional Hospital, Accra, Ghana assessed (1) the Callascope's ability to visualize the cervix compared to the standard-of-care speculum exam, (2) the feasibility and willingness of women to use the Callascope for self-exams, and (3) the feasibility and willingness of clinicians and their patients to use the Callascope for clinician-based examinations. Cervix visualization was comparable between the Callascope and speculum (83% or 44/53 women vs. 100%) when performed by a clinician. Visualization was achieved in 95% (21/22) of women who used the Callascope for self-imaging. Post-exam surveys indicated that participants preferred the Callascope to a speculum-based exam. Our results indicate the Callascope is a viable option for clinician-based and self-exam speculum-free cervical imaging.Clinical study registration ClinicalTrials.gov https://clinicaltrials.gov/ct2/show/record/ NCT00900575, Pan African Clinical Trial Registry (PACTR) https://www.pactr.org/ PACTR201905806116817.

Combining multiple contrasts for improving machine learning-based classification of cervical cancers with a low-cost point-of-care Pocket colposcope.

We apply feature-extraction and machine learning methods to multiple sources of contrast (acetic acid, Lugol's iodine and green light) from the white Pocket Colposcope, a low-cost point of care colposcope for cervical cancer screening. We combine features from the sources of contrast and analyze diagnostic improvements with addition of each contrast. We find that overall AUC increases with additional contrast agents compared to using only one source.

Evaluation of a low-cost, low-power syringe pump to deliver magnesium sulfate intravenously to pre-eclamptic women in a Malawian referral hospital.

BACKGROUND: Magnesium sulfate is an affordable and effective treatment for pre-eclampsia and eclampsia. In settings where infusion pumps are not available to regulate the flow rate of intravenous delivery, healthcare providers must administer magnesium sulfate (MgSO4) via time-consuming and painful, large-volume intramuscular injections. As an alternative to costly commercially available syringe pumps, we developed AutoSyp, an accurate, low-cost, and low-powered syringe pump designed to meet the needs and constraints these low-resource settings. This paper describes results of a pilot study to evaluate the feasibility of using AutoSyp to administer MgSO4 intravenously to women suffering from pre-eclampsia at a referral hospital in Blantyre, Malawi. METHODS: AutoSyp was programmed to deliver MgSO4 following the Zuspan regimen to pregnant and post-partum women suffering from pre-eclampsia at Queen Elizabeth Central Hospital in Blatnyre, Malawi. Given the selection of either loading or maintenance dose on AutoSyp's user interface, the flow rate was automatically programmed to dispense 60 mL/h or 5 mL/h of 20% MgSO4 solution, respectively. During each treatment, the dispensed volume was automatically calculated by the device based on the plunger position and stored on a computer for accuracy analysis of the mean flow rate and total volume delivered. The clinical results for both the loading and maintenance dose administrations were compared to the device's accuracy during tests performed in the laboratory setting. RESULTS: Twenty-two women were enrolled in this study. In both the clinical and laboratory settings, the mean flow rate errors for the loading and maintenance dose infusions were under 2%. During 466 h of testing, the device sounded 129 occlusion alarms across 14 subjects. Of these, 71 alarms were false positives. CONCLUSION: Results of this study support the use of AutoSyp as a less painful and accurate means of MgSO4 administration in clinical environments that lack infusion systems. There were a large number of false alarms in the current system which will be addressed in future designs. AutoSyp maintains the comfort of intravenous MgSO4 administration, but unlike commercially available syringe pumps, it is capable of operating with a variety of syringe brands and sizes and requires no additional consumables. AutoSyp's appropriate design will benefit its implementation and sustained use in low-resource settings. TRIAL REGISTRATION: Trial registered prospectively on November 18, 2014 with ClinicalTrials.gov ( NCT02296931 ).

AutoSyP: A Low-Cost, Low-Power Syringe Pump for Use in Low-Resource Settings.

This article describes the design and evaluation of AutoSyP, a low-cost, low-power syringe pump intended to deliver intravenous (IV) infusions in low-resource hospitals. A constant-force spring within the device provides mechanical energy to depress the syringe plunger. As a result, the device can run on rechargeable battery power for 66 hours, a critical feature for low-resource settings where the power grid may be unreliable. The device is designed to be used with 5- to 60-mL syringes and can deliver fluids at flow rates ranging from 3 to 60 mL/hour. The cost of goods to build one AutoSyP device is approximately $500. AutoSyP was tested in a laboratory setting and in a pilot clinical study. Laboratory accuracy was within 4% of the programmed flow rate. The device was used to deliver fluid to 10 healthy adult volunteers and 30 infants requiring IV fluid therapy at Queen Elizabeth Central Hospital in Blantyre, Malawi. The device delivered fluid with an average mean flow rate error of -2.3% ± 1.9% for flow rates ranging from 3 to 60 mL/hour. AutoSyP has the potential to improve the accuracy and safety of IV fluid delivery in low-resource settings.

Maji: a new tool to prevent overhydration of children receiving intravenous fluid therapy in low-resource settings.

We designed and evaluated the accuracy and usability of a device to regulate the volume of fluid dispensed during intravenous drip therapy. The mechanical system was developed in response to a pressing need articulated by clinicians in pediatric wards throughout sub-Saharan Africa, who require a tool to prevent overhydration in children receiving intravenous fluid in settings that lack burettes or electronic infusion pumps. The device is compatible with most intravenous bags and limits the volume dispensed to a preset amount that can be adjusted in 50 mL increments. Laboratory accuracy over a range of clinically-relevant flow rates, initial bag volumes, and target volumes was within 12.0 mL of the target volume. The ease of use is "excellent," with a mean system usability score of 84.4 out of 100. Use of the device limits the volume of fluid dispensed during intravenous therapy and could potentially reduce the morbidity and mortality associated with overhydration in children receiving intravenous therapy.