ISSN 2756-3391
African Journal of Parasitology Research | Vol. 14, No. 8, August 2026 | pp. 557–564
DOI: 10.46882/AJPR/170543
Article Type: Short Communication
Title: Diagnostic validation of an artificial intelligence-driven smartphone microscopic system for rapid field detection of soil-transmitted helminths in rural KwaZulu-Natal
Names of Authors: Sibusiso K. Ndlovu¹, Thabo M. Khumalo¹, Elena R. Thompson²
Authors’ Affiliations: ¹Department of Parasitology, Faculty of Health Sciences, University of KwaZulu-Natal, Durban, South Africa. ²Centre for Diagnostic Innovation, Digital Health Institute, Johannesburg, South Africa.
Abstract: Point-of-care digital diagnostics offer automated alternatives to manual microscopy in resource-constrained regions. This study evaluated the diagnostic accuracy, sensitivity, and specificity of an artificial intelligence-driven smartphone microscopic attachment (SmartMicro-AI) optimized for the automated identification and quantification of soil-transmitted helminth ova. A diagnostic validation trial was conducted using 650 stool specimens from school children in rural KwaZulu-Natal, South Africa, utilizing dual-investigator manual Kato-Katz microscopy as the reference standard. The SmartMicro-AI system uses a custom deep-learning convolutional neural network to scan thick smears via a mobile phone camera lens. The automated system achieved an overall sensitivity of 93.4 percent and a specificity of 97.2 percent for detecting Ascaris lumbricoides, Trichuris trichiura, and hookworm species combined. Species-specific analysis revealed optimal performance for Ascaris (sensitivity 96.1 percent), while hookworm sensitivity was slightly lower at 88.5 percent due to rapid egg degradation in older samples. Quantification correlation analysis showed strong agreement between AI-calculated eggs per gram values and manual counts (r = 0.89, P < 0.001). Processing time per sample was reduced from an average of 8.5 minutes manually to 1.2 minutes using the automated system. These field validation findings demonstrate that automated mobile microscopy provides reliable, scalable diagnostic outputs, offering a robust tool for large-scale epidemiological monitoring and verification of helminth elimination benchmarks.
Keywords: Soil-transmitted helminths, Digital diagnostics, Artificial intelligence, Mobile health, Point-of-care microscopy, South Africa
Manuscript Timeline: Received: 20 May 2026; Revised: 22 June 2026; Accepted: 15 July 2026; Published: 12 August 2026
Subscribe to read the full article: https://internationalscholarsjournals.org/subscribe-to-read