Back

Serological Markers Predict Plasmodium vivax Relapses in Returning Indonesian Soldier Cohorts

Noviyanti, R.; Setya Utami, R. A.; Smith, L.; Trianty, L.; Ekawati, L.; Sutanto, E.; Amalia, R.; Amelia, A. R.; Hafidzah, M. A.; Fadila, N.; Puspitasari, A. M.; Nisa, F. A.; Hidar, H.; Kariodimedjo, P.; Farinisia, A.; Hutahaean, G.; Christian, M.; Kesuma, T. A.; Subekti, D.; Soebianto, S.; Wulandari, F.; Nuraeni, N.; Budiman, W.; Ertanto, Y.; Widiarta, M. D.; Furkan, F.; Nekkab, N.; Mazhari, R.; White, M.; Robinson, L.; Longley, R.; Baird, J. K.; Mueller, I.

2026-06-10 infectious diseases
10.64898/2026.06.08.26355218 medRxiv
Show abstract

Summary Background Persistent transmission from relapsing Plasmodium vivax infections threatens malaria elimination programs in the Asia-Pacific and Americas. Tools to identify people at risk of relapse are urgently required. We aimed to validate a panel of eight P. vivax serological biomarkers for predicting future relapses. Methods In this observational study, soldiers returning from malaria-endemic Papua to non-endemic East Java, Indonesia, were screened at enrolment using antibody measurement (Luminex) and trained random forest classification algorithms, then followed for 6 months. Active case detection was performed fortnightly by microscopy. Algorithms classified soldiers as recently infected (last nine months) and thus at risk of relapse, based on anti-vivax antibody measurements at enrolment. Findings Between December 2018 and July 2022, 592 soldiers were enrolled, with 553 completing follow-up; 119 experienced a P. vivax relapse. Of these, 102 were correctly classified as at risk of relapse at enrolment, corresponding to 86% sensitivity and 86% specificity, with an AUC of 0.92. Interpretation P. vivax serological biomarkers can identify people at risk of relapse with high sensitivity and specificity and could be used as a novel public health intervention, P. vivax serological testing and treatment (PvSeroTAT), to reduce relapse-driven transmission.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.