Molecular diagnostics for cutaneous leishmaniasis: progress towards fulfilling the WHO target product profile Authors: Jan Jarzabek, Paul William Denny Published: 2026 ABSTRACT Recently, the WHO published a Target Product Profile for a diagnostic test for cutaneous leishmaniasis (CL) and a Roadmap to 2030 for Neglected Tropical Diseases. The documents highlight that existing diagnostic tools for CL are insufficient, whilst setting clear goals for improved sensitivity and reduced cost. The need for species typing in diagnostics is also becoming more pressing with the emergence of drug-resistance, especially of Leishmania tropica. Serological tests are unable to do this, while techniques that can, like PCR, require complex and expensive machinery. Isothermal assays like LAMP offer a promising solution, but more work also remains, as few species-specific LAMP assays have been developed thus far and CL in Ethiopia is particularly neglected. Additionally, since the COVID-19 pandemic, many cheap isothermal diagnostic devices have been produced, which have yet to be tested in the diagnosis of CL. Finally, artificial intelligence presents another avenue for rapid diagnosis by image analysis. In this comprehensive review, we examine the opportunities and challenges inherent to diagnostic development for CL, a priority undertaking that still faces many developmental hurdles. Introduction Leishmaniasis is one of 25 neglected tropical diseases (NTDs) characterised by their impact on 100’s of millions of the global poor and a chronic lack of attention and investment (World Health Organisation, 2021). Caused by the protist Leishmania spp., the disease has three primary clinical presentations: cutaneous (CL), mucocutaneous (MCL) and visceral (VL) (Georgiadou et al. 2015). CL involves lesions on the skin and accounts for the majority of cases (Aronson et al. 2016). MCL features infection of the mucous membranes of the respiratory tract, while VL is an infection of the internal organs such as the liver and spleen (Aronson et al. 2016). Currently, 49% countries reporting to the WHO are considered endemic for leishmaniasis, 45% for CL specifically (Jain et al. 2024b). The most recent WHO figures (2023) report 272 000 cases of CL, with the vast majority (81%) occurring in the Eastern Mediterranean Region (Figure 1) (Jain et al. 2024b). This is a slight reduction from the peak of 281 000 cases in 2019; however, there has been an upward trend since 2014 (Jain et al. 2024b). At the same time, VL case numbers fell from 31 000 to 12 000 (Jain et al. 2024b). Another study reported a 95% reduction in VL incidence globally from 1990 to 2021 (Zhang et al. 2025c). This positive trend is due to public health interventions focused on VL, most notably the joint eradication programme in India, Bangladesh and Nepal (Thakur et al. 2020; Jain et al. 2024b). In fact, VL is one of eight NTDs which have been eliminated in at least one country, in this case Bangladesh (Hietanen et al. 2025). Unfortunately, CL remains neglected – a recent review found that the total burden of CL across North Africa and the Middle East had not changed from 1990 to 2021 (EL Banna et al. 2025). It has also been suggested that, due to the shortage of diagnostic tools, CL case numbers are underreported and could be as high as one million per annum (de Vries and Schallig, 2022).Figure 1.The Eastern Mediterranean Region (EMR), as defined by the WHO, comprising Afghanistan, Bahrain, Djibouti, Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Pakistan, Palestine, Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tunisia, UAE, and Yemen.Figure 1 long description.A map highlighting countries in the Eastern Mediterranean Region as defined by WHO. See long description. The Eastern Mediterranean Region (EMR), as defined by the WHO, comprising Afghanistan, Bahrain, Djibouti, Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Pakistan, Palestine, Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tunisia, UAE, and Yemen. Figure 1 long description. The WHO NTD Road Map to 2030 identifies diagnostics as an area in CL research where critical action is required, with a need to improve the sensitivity and affordability and a goal of 85% of cases being detected in every endemic country by 2030 (World Health Organisation, 2021). A recent report concluded that these goals had not yet been met (Donadeu et al. 2025). Indeed, based on the latest figures from the WHO, only 27% of cases are being reported (Jain et al. 2024b). Challenges in CL diagnosis which may account for this failure include the relative lack of resources and medical infrastructure, political instability and the presence of multiple endemic species (Akhoundi et al. 2017; de Vries and Schallig, 2022). Additionally, CL lesions can be misdiagnosed as a number of communicable and non-communicable diseases such as cutaneous tuberculosis, psoriasis, purulent dermatitis and pyoderma gangrenosum due to similarities in appearance (Di Altobrando et al. 2021; Sikorska et al. 2022; Karaja et al. 2024; Shrestha et al. 2024). To address these challenges, a target product profile (TPP) by the Foundation for Innovative Diagnostics, in collaboration with the Drugs for Neglected Diseases initiative and the WHO outlines the desired characteristics for a CL diagnostic test (Cruz et al. 2019). The TPP lists criteria such as specificity (95–100%), sensitivity ( > 90%) and ease of use, both with respect to performing the test and in sample collection and preparation. The TPP also lists the need for species-level discrimination and the ability to withstand diverse and extreme storage conditions such as heat, humidity and the lack of a cold chain. Finally, the document suggests a cost per test of $1–5, with a testing device costing ≤ $2000. The goals of the TPP overlap with the WHO-coined acronym REASSURED which is a general guideline for a successful diagnostic test: Real-time connectivity, Ease of specimen collection, Affordable, Sensitive, Specific, User-friendly, Rapid and robust, Equipment-free and Deliverable to end-users (Otoo and Schlappi, 2022). In this review, we discuss the shortcomings of existing CL diagnostics, as well as recent developments which may enable the development of a diagnostic tool matching the benchmarks in the WHO TPP by the 2030 deadline. Aetiology and pathogenesis of CL Human leishmaniasis is caused by 20 different species of Leishmania (Akhoundi et al. 2016); the most common disease-causing ones are summarised in Figure 2. Old World CL is caused by species of the Leishmania (Leishmania) subgenus, namely Leishmania major, Leishmania tropica, Leishmania aethiopica and less frequently, Leishmania infantum and Leishmania donovani (Akhoundi et al. 2016; Gurel et al. 2020). New World CL and MCL are caused by L. (Viannia) species such as Leishmania braziliensis, Leishmania guyanensis and Leishmania panamensis, although L. (Leishmania) species such as L. infantum, L. mexicana and Leishmania amazonensis are also endemic in the region (Akhoundi et al. 2016; Gurel et al. 2020).Figure 2.The most common Leishmania species causing disease in humans. Generated using BioRender.Figure 2 long description.Diagram showing Leishmania species distribution between Old and New World. See long description. The most common Leishmania species causing disease in humans. Generated using BioRender. Figure 2 long description. The disease is transmitted to the mammalian host via a bite from the insect sandfly vector (Akhoundi et al. 2016). The simplest clinical presentation is localised CL: individual self-healing lesions at the site of infection (Aronson et al. 2016). Leishmania major infections typically follow this pattern, with lesions usually healing after 2–6 months (Auwera G and Dujardin, 2015). In contrast, L. tropica lesions typically take 6–15 months to heal (Aronson et al. 2016). Leishmania major CL was first distinguished by its ‘wet rural’ lesions, as opposed to the ‘dry urban’ lesions caused by L. tropica (Akhoundi et al. 2016). Leishmania tropica is traditionally considered anthroponotic (human to vector to human transmission), as opposed to most other Leishmania infections which are zoonotic (animal to vector to human transmission) (Gurel et al. 2020; Reimão et al. 2020). This urban, anthroponotic transmission pattern makes L. tropica a great threat in times of large population displacement, such as during the Syrian civil war or the recent regime change in Afghanistan (Kim et al. 2025; Rahimi et al. 2025). L. tropica is also considered more genetically diverse than L. major (Schwenkenbecher et al. 2006; Pratlong et al. 2009; Iantorno et al. 2017). Leishmania aethiopica is a CL-causing species found exclusively in Ethiopia and neighbouring countries (Blaizot et al. 2024). Ethiopia is 1 of the 10 most affected countries for CL, with an estimated 20 000 to 50 000 cases annually (Taye et al. 2024). Although there is some discrepancy in recent studies about the level of genetic diversity, it was indicated that L. aethiopica is more diverse than expected given its limited range (Hadermann et al. 2023; Yizengaw et al. 2024). This diversity may account for the poor performance of commercial diagnostic tests reported in Ethiopia (Van Henten et al. 2022; Taye et al. 2024). As mentioned above, L. infantum and L. donovani can cause CL despite being primarily viscerotropic (Akhoundi et al. 2016). L. donovani can also cause post-kala-azar dermal leishmaniasis (PKDL) where, a visceral disease is followed by dermal symptoms (Reimão et al. 2020). These atypical disease forms are a challenge to VL elimination efforts as they create an additional reservoir of L. donovani (Jain et al. 2024a; Bhattarai et al. 2025). Other complex forms of CL include leishmaniasis recidivans caused by L. tropica and diffuse CL (DCL) and MCL caused by L. aethiopica (Krayter et al. 2014, 2015; Aronson et al. 2016; Atnafu et al. 2024). A recent study in Ethiopia found a lack of consensus in how different physicians classify DCL and MCL (van Henten et al. 2025). This also poses a challenge as systemic treatment is recommended for these atypical CL forms, as opposed to localised CL. New World CL species such as L. amazonensis, L. braziliensis, L. guyanensis and L. panamensis also present a complex picture as they can cause both CL and MCL, while only Leishmania mexicana causes CL exclusively (Mann et al. 2021). L. infantum, often named Leishmania chagasi in the American context, also causes New World CL and VL but not MCL (de Vries and Schallig, 2022). Whilst self-healing lesions can be left untreated, incorrect diagnosis and treatment can result in scarring and disfigurement, as well as opportunistic secondary infections (Akuffo et al. 2023; Aminizadeh et al. 2024; Jaimes et al. 2024; Al-Alousy and Al-Nasiri, 2025). There are no universally efficacious treatments for CL as some Leishmania species in certain geographical settings show drug resistance (Aronson et al. 2016). For example, antimonial resistance in L. aethiopica and L. tropica infections has been reported in Ethiopia, Iran and Pakistan (Kämink et al. 2021; Solomon et al. 2022; Zewdu et al. 2022). Drug resistance has also been reported in L. mexicana and L. braziliensis (Ahmed et al. 2020; Gonzalez-Garcia et al. 2025; Ko et al. 2025; Moreno-Rodríguez et al. 2025). For the reasons described above, species identification is crucial to the effective treatment and control of CL. Unfortunately, due to the costly and time-consuming nature of species typing, data on the prevalence of individual CL Leishmania species remains limited and sporadic (Figure 3). A study in the Khyber Pakhtunkhwa province in Northern Pakistan found that 90% of CL cases were caused by L. tropica, despite the initial clinical examination finding that 73% of the lesions were dry (presumed L. tropica) and 27% were wet (presumed L. major) (Ullah et al. 2024). This indicated that a majority of the 118 patients who had wet lesions were actually infected with L. tropica, which is an atypical presentation and demonstrates that clinical symptoms alone are insufficient for accurate diagnosis. A nationwide study found that 79% of CL cases were L. tropica, 15% were L. infantum and 6% were L. major, confirming that L. tropica is the main driver of CL in Pakistan (Naz et al. 2024). In neighbouring Afghanistan L. tropica has historically accounted for the majority of cases, although more recent data on this is not available (Rahimi et al. 2025). In Israel, Mauritania and Niger, the situation is reversed, with 84–100% of cases caused by L. major (Avni et al. 2024; Blaizot et al. 2025; El Moctar et al. 2025). In Iran the dominant species can be L. major or L. tropica, depending on the region (Zeinali et al. 2023; Hayatolgheib-Moghadam et al. 2025; Marvi-Moghadam et al. 2025). In Morocco 52% of cases were caused by L. tropica, 39% L. major and 9% L. infantum (Baghad et al. 2025). Recent studies in Ethiopia found L. aethiopica in 96% to 100% of samples (Amare et al. 2023; Tesfaye et al. 2025). Interestingly, a study of a recent CL outbreak in the conflict-stricken Somali region of Ethiopia, reported 100% of samples as L. tropica (Abera et al. 2025). However, this finding should be treated with a note of caution as only 18 of the 1050 patients in the study were sampled for species typing. Nevertheless, in 99% of all 1050 patients the lesions examined were wet, unlike the typically dry lesions caused by L. tropica or L. aethiopica. The outbreak was also characterised by an unusually high number of lesions per patient, with 90% of patients having seven or more lesions.Figure 3.Countries where OWCL species prevalence was reported in the years 2023–2025.Figure 3 long description.A thematic map of North Africa, the Middle East and South Asia showing labeled Leishmania species percentages for Morocco, Mauritania, Niger, Israel, Iran, Pakistan and Ethiopia; north is at top. See long description. Countries where OWCL species prevalence was reported in the years 2023–2025. Figure 3 long description. Equivalent data for New World CL is also sparse. A study in Roraima state in Brazil found L. braziliensis in 32% of samples, L. amazonensis in 17% of samples and L. panamensis in 13% of samples (de Almeida et al. 2021). A further 6 species were found in 14% of samples and in 24% of samples the species could not be identified. A study in Colombia found L. panamensis in 77% of samples, L. guyanensis in 19% of samples and L. braziliensis in 4% of samples (Hoyos et al. 2022). In a study in Bolivia, 50% of CL samples were identified as L. braziliensis and 50% as subpopulation called L. braziliensis outlier (Torrico et al. 2022). Finally, a study in Panama found L. panamensis in 59% of samples and L. guyanensis in 41% of samples, although not all results could be confirmed after further testing (Reina et al. 2022). Traditional diagnostic methods As discussed, CL lesions can be visually classified as wet or dry, with additional criteria such as nodular, popular, plaque or ulcerated (Naz et al. 2024; Ullah et al. 2024). As mentioned, these characteristics correlate with the infecting species, but not reliably enough for accurate diagnosis (Naz et al. 2024; Ullah et al. 2024). CL can be confirmed by microscopic examination of a direct smear from the lesion (de Vries and Schallig, 2022). A Giemsa stain allows the Leishmania amastigote forms (the mammalian stage) to be identified, however is not species specific, requires technical expertise and has limited sensitivity (50–70%) (Aronson et al. 2016; Reimão et al. 2020; de Vries and Schallig, 2022). Parasite isolation by culture can improve sensitivity and allows further analytical tests such as PCR, but it prolongs the diagnosis time, is highly technical and liable to contamination (Reimão et al. 2020). Patient sampling methods include scraping, brushing, swabbing and aspiration (Reimão et al. 2020). The choice of technique is influenced by cosmetic significance and patient comfort (Aronson et al. 2016). A recent study compared methods for cutaneous lesions – skin slit, dental broach, tape dis and microbiopsy (van Henten et al. 2024). The skin slit had the highest self-reported pain score with a median of 6/10 compared to dental broach (4/10) and tape disc and microbiopsy (both 1/10). Tape disc and microbiopsy had poorer sensitivity (69–83%), compared to the dental broach (96%) and skin slit (95%). Another study comparing lancet scraping, filter paper imprints and microbiopsy samples for kDNA-PCR found that lancet scraping was the most sensitive method (99%) compared to filter paper imprints (95–96%) and microbiopsy (93–94%) (De Los Santos et al. 2024). Serological and immunological tests Serological and immunological tests like the Leishmanin Skin Test have been in use for almost 100 years (Dey et al. 2023). The direct agglutination test (DAT), a later development where killed Leishmania promastigotes are mixed with patient blood or serum, was found to have 95% sensitivity and 95% specificity for VL in a recent meta-analysis (Roberts et al. 2023). Commercial ELISA tests for VL are also available. A recent study compared NovaLisa (Novatec), Bordier (Bordier Affinity), Ridascreen (R-Biopharm) and Vircell (Vircell) (Lévêque et al. 2020). The tests showed 81–94% sensitivity and 96–100% specificity, although they take 2.5–3.5 hours to perform. The four commercial ELISAs also performed poorly on samples from immunocompromised patients, with 15–42% false negative results. A separate study reported 96% sensitivity and 81% specificity for NovaLisa (Rune Stensvold et al. 2019). Despite the promising performance with VL samples, these kinds of tests are not recommended for CL as the infection does not produce a strong humoral response (de Vries and Schallig, 2022). Additionally, serological tests do not distinguish past and active infection, and cross-reactivity has been shown with Chagas Disease, caused by the closely related Trypanosoma cruzi, but also with leprosy, tuberculosis, typhoid fever and malaria (Aronson et al. 2016). Despite this, a meta-analysis of patents for Leishmania diagnostics in the years 2010–2022 found that 63 of the patents were based on immunological methods whereas only 33 were based on molecular methods (de Avelar et al. 2023). Among those 33, only six were tested against L. tropica and none against L. aethiopica respectively. PCR-based methods PCR is an established laboratory technique for amplifying a specific DNA sequence with end-point or real-time detection (qPCR) (Auwera G and Dujardin, 2015). Downstream analyses, such as restriction fragment length polymorphism (RFLP) or nested PCR can improve specificity and sensitivity (Akhoundi et al. 2017). Both approaches have been used for species identification in Leishmania spp. (Ashraf et al. 2025; Mohammadi Manesh et al. 2025). Melting curve analyses can also be used in this way, as the melting temperature of double-stranded DNA is sequence-specific. The technique has been used to differentiate PCR products from L. major, L. tropica and L. donovani (Azam et al. 2024). Diagnosis of leishmaniasis by PCR-based approaches has already been reviewed extensively (Auwera G and Dujardin, 2015; Akhoundi et al. 2017; Reimão et al. 2020; Thakur et al. 2020; Nharwal et al. 2025). One of the most common targets is the ribosomal gene array, particularly the 18S subunit gene and the intergenic spacer ITS1 (Boulal et al. 2025; Mughal et al. 2025). The kDNA minicircle, a DNA fragment of approximately 800bp found in the kinetoplast has also been used (Roozbehani et al. 2025). The kinetoplast is a DNA-containing structure in the mitochondria that is unique to the Kinetoplastida, such as Leishmania and Trypanosoma spp. and has long been favoured as a PCR target due to its very high copy number reaching as many as 10 000 copies (Noyes et al. 1998). The kinetoplast also comprises larger fragments called maxicircles containing the cytochrome b gene, which is another target (Mughal et al. 2025). Recently, cytochrome c oxidase I, also located on the maxicircle, has been identified as a new alternative (Mata-Somarribas et al. 2024). Others include cysteine protease B (cpb), glycoprotein GP63, heat-shock protein 70 (hsp70) and 7SL-RNA (Chaouch et al. 2013; Hosseini-Safa et al. 2018; Azam et al. 2024). A recent meta-analysis covering the years 2011–2022 found a pooled sensitivity of 91% and pooled specificity of 98% for qPCR (Rihs et al. 2025b). An update by the same authors found that sensitivity and specificity of a qPCR approach against VL was 90% and 99.6% overall, but 86% and 100% when non-invasive sampling was used (Rihs et al. 2025a). The authors cite urine samples as a non-invasive alternative to blood and bone marrow samples which are typically used in the diagnosis of VL. For CL, the sensitivity and specificity were 87% and 87% when invasive sampling was used, but only 70% and 96% using non-invasive sampling. Urine samples and lesion swabs, imprints and aspirates were classified as non-invasive methods as opposed to lesion biopsies, scrapings and smears. The authors note that for VL samples the sensitivities reported in the original papers were often calculated using serological tests as the reference (Touria et al. 2018; Iatta et al. 2021; Pereira et al. 2021). For CL samples, because of the weaker humoral response the reference techniques used were microscopy, immunohistochemistry or a second molecular test such as conventional PCR (Vink et al. 2018; Chaouch et al. 2019; Silgado et al. 2021). Therefore, while the gene targets are often the same, comparisons in sensitivity between VL and CL diagnostic assays should be made with caution due to these discrepancies. Despite the promising results of various PCR-based assays in the laboratory, their deployment in the field is hampered by the requirement for expensive laboratory facilities and trained technical staff (Thakur et al. 2020). Furthermore, the reduced sensitivity to CL compared to VL, particularly when non-invasive sampling is used, is a notable drawback. In an attempt to overcome the technical barrier, miniature PCR devices have been developed. Palm PCR, a miniature battery-powered PCR incubator, was able to detect L. tropica, L. infantum and L. major with a limit of detection of 40 pg, 4 pg and 0.4 pg, respectively (Bel Hadj Ali et al. 2023). The authors created a duplex assay which could distinguish L. major, L. tropica and L. infantum and by coupling the PCR reaction to a lateral flow test (LFT) the results could be visualised in under 30 minutes. One drawback of the study was the lack of clinical samples tested. An alternative device, the miniPCR thermocycler, has been used to detect L. panamensis (Castellanos-Gonzalez et al. 2023). The authors used the Prismo Mirage smartphone app to quantify the fluorescent signal obtained with the miniPCR transilluminator. Amplification by miniPCR had equal sensitivity (100%) to a reference qPCR assay when tested on samples from Colombia and Peru, although the time to amplification was not recorded. The miniPCR has also been used in combination with an LFT against VL samples from Ethiopia (Hagos et al. 2024). The test showed 96% sensitivity and 99% specificity relative to qPCR, giving results in just over an hour. The same assay was tested on VL samples from Spain and East Africa (van Dijk et al. 2024). In Spain, miniPCR had 96% sensitivity and 97% specificity, while in East Africa it had 88% sensitivity and 100% specificity when compared to nested PCR. A few commercial PCR kits have been developed for the diagnosis of leishmaniasis; a recent study compared STAT-NAT (Sentinel), Real-TM (Sacace) and VIASURE (CerTest) (Arnau et al. 2023). Whilst Real-TM requires −20oC storage, STAT-NAT and VIASURE both allow storage at room temperature. When tested against a range of species, VIASURE and Real-TM detected both Old World and New World species, while STAT-NAT could only detect Old World species and was therefore excluded from further analysis. When tested against clinical samples, VIASURE had 82% sensitivity and 100% specificity compared to Real-TM. Another interesting development has been adaptive PCR, where fluorescently labelled L-DNA probes are used to model primer annealing efficiency, which allows the PCR conditions to be adjusted in real time and the amplification rate to be increased accordingly. A recent study using this technique was able to achieve 20 cycles of amplification in under 15 minutes due to the reduced cycle duration (Spurlock et al. 2025). An alternative method is microfluidic qPCR, such as that developed by GeneSoC, which was shown to be able to achieve 50 cycles in 15 minutes (Mizushina et al. 2025). These findings, which have not been applied to Leishmania yet, could enable the development of a cheap and rapid PCR test for leishmaniasis for use in the field. Isothermal amplification Loop-mediated Isothermal Amplification (LAMP) was developed in 2000 (Notomi et al. 2000) and uses a thermostable Bst Polymerase with strand displacement activity. This, together with the use of a second outer pair of primers, facilitates strand separation without the high temperature melting step used in PCR and allows the whole reaction to take place at the same temperature (isothermal). A subsequent adaptation incorporating a third primer pair increases the amplification speed (Nagamine et al. 2002). Various approaches for monitoring the LAMP reaction, such as colorimetric or fluorescent dyes and turbidity caused by the byproduct pyrophosphate, are available (Figure 4) (Poirier et al. 2023; Jin et al. 2024; Quero et al. 2025). The assay can also be coupled with LFTs for ease of detection (Sanmoung et al. 2023). LAMP-LFT has not been used in the diagnosis of leishmaniasis, but similar approaches combining LFTs with PCR and RPA (Recombinase Polymerase Amplification, an alternative isothermal technique) have been reported (Cossio et al. 2021; Hagos et al. 2024; Nawattanapaibool et al. 2024; van Dijk et al. 2024).Figure 4.A visual summary of common approaches and devices used for the detection of LAMP amplification. Generated using BioRender.Figure 4 long description.Labeled scientific infographic comparing LAMP amplification detection methods and readouts, including turbidity, dyes, molecular beacons, CRISPR, lateral flow strip, microfluidic chip and thermal cycler. See long description. A visual summary of common approaches and devices used for the detection of LAMP amplification. Generated using BioRender. Figure 4 long description. The simultaneous use of three primer pairs provides a high degree of specificity. Indeed, LAMP assays have been used to detect point mutations conferring drug resistance (Sanmoung et al. 2023). Specificity can be further improved by detection with molecular beacons. A molecular beacon is a fluorophore in a quenched form, either as dsDNA or hairpin ssDNA. The fluorophore is released from the quencher when the beacon binds a specific sequence on the LAMP amplicon (Xue et al. 2024). This approach allows multiplexing as different fluorophores can be used with each beacon (Sherrill-Mix et al. 2021). The technique has not been applied to Leishmania spp., but it has been tested with Trypanosoma brucei and Schistosoma mansoni (Wan et al. 2017; Crego-Vicente et al. 2023). An alternative approach for increasing specificity is integration with CRISPR/Cas9 (Liang et al. 2024). Briefly, the Cas endonuclease becomes activated by binding a target sequence on the LAMP amplicon with the help of a guide RNA and, subsequently, cleaves a short reporter DNA fragment to release the fluorophore from its quencher. The specific combination of LAMP and CRISPR has not been applied to leishmaniasis yet, although other studies have combined CRISPR with PCR and with RPA (Dueñas et al. 2022; Yang et al. 2024). CRISPR-assisted LAMP has also been used to diagnose toxoplasmosis, demonstrating its feasibility in parasitic infections (Liang et al. 2024). In summary, the high sensitivity and specificity due to the use of three primer pairs, the rapid amplification time and the simple technical requirement of an isothermal heat source, make LAMP a very promising technique for a low-cost, species-specific, field diagnostic, such as that required for CL. Multiple LAMP assays for the diagnosis of leishmaniasis have been published and reviewed previously (Auwera G and Dujardin, 2015; Akhoundi et al. 2017; Nzelu et al. 2019; Reimão et al. 2020; Thakur et al. 2020; de Vries and Schallig, 2022; Nharwal et al. 2025). The gene targets used are typically the same as those for PCR as they share the same characteristics, namely interspecies variation and high copy number. The most recent novel assay targeting hsp70 achieved 100 fg sensitivity (Soares et al. 2024). This is a useful benchmark as it is the approximate amount of DNA in a single Leishmania parasite cell (Soares et al. 2024). The study found no cross-reactivity to T. cruzi and T. brucei, which had been observed with other LAMP assays for Leishmania spp. However, the reaction time of 60 minutes is a major drawback. Furthermore, the test is not species specific and the study focused on New World Leishmania spp. such as L. braziliensis, and whilst it includes some Old World species like L. major and L. infantum, it does not include L. tropica and L. aethiopica. Diagnostic assays for leishmaniasis often focus on species for which effective diagnostic tests are already available, like L. donovani and L. infantum, while species which cause a substantial part of the CL burden like L. tropica and L. aethiopica remain overlooked. These choices are undoubtedly informed by existing knowledge of endemic species in the regions where researchers are based, but this is by no means fixed in stone. Recent years have shown that Leishmania species can be rapidly introduced into non-endemic countries, either by human migration or by animal reservoirs (Azami-Conesa et al. 2024; Kim et al. 2025). To date, only three studies have designed LAMP primers to discriminate Leishmania at the species level. The first used primers targeting the kDNA minicircle of L. donovani with a limit of detection of 1 fg and no cross-reactivity against L. infantum, L. major or L. tropica (Verma et al. 2013). A further study used primers targeting CPB to distinguish L. major and L. tropica with a limit of detection of 20 fg and 200 fg, respectively (Chaouch et al. 2019). Most recently, one study used primers targeting kDNA minicircle of L. tropica with a limit of detection of 1 fg and no cross-reactivity for L. major or L. infantum (Taslimi et al. 2023). All of these assays demonstrated detection levels of two cell equivalents or less. RPA is an alternative isothermal technique which requires lower temperatures (37oC–42oC) than LAMP (Wu et al. 2023). However, the technique requires a recombinase and a single-strand DNA-binding protein in addition to the DNA polymerase which may be a challenge with respect to developing a lyophilisation protocol. Nevertheless, some recent studies have successfully used RPA to detect Leishmania spp. (Danthanarayana et al. 2023; Ghosh et al. 2023; Roy et al. 2023; Wu et al. 2023; Bel Hadj Ali et al. 2024; Kobialka et al. 2024; Wiggins et al. 2024). Sequencing-based approaches Targeted and whole-genome DNA sequencing (WGS) has promising uses in diagnostics. Species typing of a CL infection by targeted sequencing using the Oxford Nanopore MinION platform has been reported and the technique can show evidence of mixed or hybrid infection (Imai et al. 2018; Maia de Souza et al. 2022; Patiño et al. 2023; Jaimes et al. 2024). The platform has likewise been used in the detection and genetic typing of other protozoan parasites such as Trypanosoma, Toxoplasma and Plasmodium spp. (Cruz-Saavedra et al. 2024; Huggins et al. 2024; Koutsogiannis and Denny, 2024). While targeted sequencing can provide an indication of mixed or hybrid infection, WGS is required to distinguish between the two, as true hybrids are marked by genome-wide changes in somy and allele frequencies (Lypaczewski et al. 2022; Van den Broeck et al. 2023). Monitoring the emergence of hybrids is crucial as these can produce atypical forms of disease (Lypaczewski et al. 2022; Bruno et al. 2024). WGS can also be used for surveillance, as was shown with L. donovani isolates in Nepal (Monsieurs et al. 2024). In the context of the current VL elimination program in Nepal, this can be used to distinguish VL cases caused by strains introduced from external sources from cases caused by endemic strains. Finally, generating new genome assemblies can provide more data on sequences and copy numbers for diagnostic targets such as hsp70, 18S and GP63 (Alonso et al. 2016; Bussotti et al. 2018). Point-of-care and rapid diagnostic tests Immunochromatographic LFTs (ICT) are widely used in the detection of VL and include IT LEISH (Bio-Rad), TruQuick (Meridian) and Kalazar Detect (InBios). In two recent studies on VL samples from Brazil and the Mediterranean, the sensitivity of these tests was 85–96% while specificity was 96–99% (Freire et al. 2019; Lévêque et al. 2020). However, the sensitivity with samples from immunocompromised patients was poorer, dropping to as little as 47%. An alternative ICT specific for CL is CL Detect (InBios), targeting the peroxidoxin antigen. A recent meta-analysis of studies testing CL Detect found 68% pooled sensitivity and 94% pooled specificity (Gebremeskele et al. 2023). The authors noted that the lack of consistency in the reference diagnostic used explained some of the discrepancies. Studies where microscopy was used as the reference reported higher sensitivity (83%) than those which included the more sensitive PCR as reference (48%). One of the studies reviewed compared different sampling methods. Sampling by dental broach had 64% sensitivity and 92% specificity, while lancet scraping had 83% sensitivity and 78% specificity (Grogl et al. 2023). One study already discussed above also compared skin slit, dental broach, microbiopsy and tape disc (van Henten et al. 2024). Skin slit and dental broach performed similarly (95–96% sensitivity), tape disc and microbiopsy samples were considerably poorer (69–83% sensitivity). Notably, when skin slit PCR was used as the reference, the other three sampling methods had very low specificity (27–73%) compared to when a composite reference was used (85–100%). As multiple authors have noted that the sensitivity and specificity of a test can be distorted depending on which technique is used as a reference, more effort is needed in future studies to ensure consistency (Van Henten et al. 2022; Gebremeskele et al. 2023). While the TPP targets for sensitivity and specificity are expressed relative to microscopy, PCR is a more sensitive technique; the ideal approach is to use both as a composite reference, making it easier to compare between studies. Loopamp is a LAMP-based test kit developed by Eiken Chemical, a spin-out company created by the authors of the LAMP technique (Feddema et al. 2024). This multiplex assay, with genus-specific primers targeting both the 18S rRNA gene and the kDNA minicircle, is lyophilised, allowing storage at room temperature. A recent meta-analysis found Loopamp had a pooled sensitivity of 96% for VL and 93% for CL, and 99% and 87% specificity, respectively (Taye et al. 2025). This makes Loopamp the best-performing diagnostic test for CL to date and highlights the potential of LAMP as the most optimal technique for fulfilling the requirements of the TPP (Table 1). Also promisingly, sensitivity and specificity were 96% and 99% when crude Boil&Spin DNA extraction was used, compared to 97% and 99% with the commercial Qiagen extraction kit. As this was tested in VL samples only, caution is needed. Additionally, as there is a variety of sampling methods for CL lesions, with no consensus on which one is most effective, much work remains to be done in this area. Table 1.The performance of various diagnostic tests for leishmaniasis as calculated by recent meta-analyses, compared to the WHO target product profileTable 1 long description.Across meta-analyses, Loopamp for VL shows the highest accuracy (96–97% sensitivity, 99% specificity), while CL Detect for CL is lowest (68% sensitivity, 94% specificity) versus the 90%/95–100% target. See long description.Test nameSensitivity (%)Specificity (%)ReferencesTarget product profile9095–100(Cruz et al. 2019)Direct agglutination test (DAT)VL9595Roberts et al. (2023)CL detect (InBios)CL6894Gebremeskele et al. (2023)qPCR–––VL9099.6Rihs et al. (2025b); Rihs et al. (2025a)CL8787–Loopamp (Eiken chemical)–––VL Qiagen DNA extraction kit9799Taye et al. (2025)VL crude DNA extraction9699–CL Qiagen DNA extraction kit9387– Table 1.The performance of various diagnostic tests for leishmaniasis as calculated by recent meta-analyses, compared to the WHO target product profileTable 1 long description.Across meta-analyses, Loopamp for VL shows the highest accuracy (96–97% sensitivity, 99% specificity), while CL Detect for CL is lowest (68% sensitivity, 94% specificity) versus the 90%/95–100% target. See long description.Test nameSensitivity (%)Specificity (%)ReferencesTarget product profile9095–100(Cruz et al. 2019)Direct agglutination test (DAT)VL9595Roberts et al. (2023)CL detect (InBios)CL6894Gebremeskele et al. (2023)qPCR–––VL9099.6Rihs et al. (2025b); Rihs et al. (2025a)CL8787–Loopamp (Eiken chemical)–––VL Qiagen DNA extraction kit9799Taye et al. (2025)VL crude DNA extraction9699–CL Qiagen DNA extraction kit9387– The performance of various diagnostic tests for leishmaniasis as calculated by recent meta-analyses, compared to the WHO target product profile Table 1 long description. Test name Sensitivity (%) Specificity (%) References Target product profile 90 95–100 (Cruz et al. 2019) Direct agglutination test (DAT) VL 95 95 Roberts et al. (2023) CL detect (InBios) CL 68 94 Gebremeskele et al. (2023) qPCR – – – VL 90 99.6 Rihs et al. (2025b); Rihs et al. (2025a) CL 87 87 – Loopamp (Eiken chemical) – – – VL Qiagen DNA extraction kit 97 99 Taye et al. (2025) VL crude DNA extraction 96 99 – CL Qiagen DNA extraction kit 93 87 – Other improvements are also needed. Loopamp showed only 48% sensitivity on CL samples in Ethiopia (Taye et al. 2024). One drawback of that study was that tape discs were used for sample collection which have been shown elsewhere to be less sensitive than other sampling methods (van Henten et al. 2024). CL Detect also had its poorest result in Ethiopia, showing 23–31% sensitivity (Van Henten et al. 2022). This indicates a broader problem with the genetic diversity of L. aethiopica being unaccounted for when these tests are being designed. Furthermore, there are no published trials of Loopamp in Pakistan and Syria, which have borne a large part of the CL burden in recent years. Testing clinical samples from a wide range of geographical locations is crucial. A recent study on diagnostic PCR assays for soil-transmitted helminth infections demonstrated that global genetic differences in primer binding sites can have a detrimental impact on assay sensitivity as these assays are often based on reference strains of the pathogen (Papaiakovou et al. 2025). Finally, despite the apparent efficacy of Loopamp outside of Ethiopia, it is only a genus-specific test. At the time of writing, there is no species-specific RDT for Leishmania spp. This feeds into a vicious cycle: because species typing in Leishmania spp. is time-consuming and requires laboratory facilities and trained staff, data on individual species prevalence is sparse, which hinders the development of future diagnostic tests. A handful of promising species-specific isothermal NAATs have been published, but further work needs to be done validating these across a variety of strains and diverse clinical samples and new assays need to be designed to cover species and geographical foci that have been previously overlooked. Isothermal amplification devices The earliest specialised device for isothermal amplification was the Genie I (Optigene), first used in a study in 2010 (Tomlinson et al. 2010). The miniaturised real-time fluorimeter was in its third iteration (the Genie III) by the time it was trialled in the COVID-19 pandemic (Fowler et al. 2021). Although prices for the Genie are not publicly available, they can be found through secondary retailers as costing $10 000–$21 000. This is cheaper than traditional real-time PCR thermocyclers, but the price still exceeds the $2000 TPP target price (Cruz et al. 2019). More recently, Eiken Chemical have developed the LF-160, an isothermal incubator with a UV transilluminator for qualitative visual detection, costing $2000 (Sohn et al. 2019). Most trials of Loopamp used the LF-160 (Mukhtar et al. 2018; Hagos et al. 2021; Hossain et al. 2021). Only one study tested Loopamp in three different devices – the LF-160, the Genie III and ESE-Quant (Qiagen) (Ibarra-Meneses et al. 2018). All three devices showed the same sensitivity, although amplification in the Genie III was much faster and matched the amplification times of the qPCR assay more closely than ESE-Quant. As the LF-160 incubator does not have a real-time function there was no equivalent data to compare. The development of isothermal diagnostic devices accelerated with the COVID-19 pandemic and many of these have been reviewed elsewhere (Das et al. 2022; Narasimhan et al. 2023). A selection of devices published since is summarised in the Supplementary Information. Most studies use fluorescent detection (54%), followed by colorimetric (20%) and electrochemical (19%) (Figure 5). Most devices were designed for PCR strips (14%) or individual PCR tubes (24%) (Figure 6). Microfluidic chips were the next most common format, being used in 29% of studies. Microfluidic devices are a promising development in diagnostics but they need to be custom-made for each application, while PCR tubes have a cheap, simple and uniform design, which provides versatility and is an advantage where low cost and ease of use are the priority.Figure 5.Detection methods (A) and sample formats (B) used in new isothermal diagnostic devices.Figure 5 long description.A pie chart showing detection methods and sample formats, with two sub-images labeled A and B and percentage ranges from 3 percent to 54 percent. See long description. Figure 6.Classes of pathogens tested in new isothermal diagnostic devices.Figure 6 long description.A pie chart showing classes of pathogens tested in new isothermal diagnostic devices, with Virus 52 percent, Bacteria 35 percent, Parasite 9 percent, Fungus 1 percent and Other 3 percent. See long description. Detection methods (A) and sample formats (B) used in new isothermal diagnostic devices. Figure 5 long description. Classes of pathogens tested in new isothermal diagnostic devices. Figure 6 long description. Only 20 (34%) of the studies listed a manufacturing cost for their device; of those, 13 (22%) were cheaper than the mini-PCR thermocycler, which costs $600 (Castellanos-Gonzalez et al. 2023). Crucially, all but one of the devices with a listed cost met the TPP target of $2000 or less. Notably, the listed devices were tested on a range of different species, primarily viruses (52%) and bacteria (35%), with parasites accounting for only 9% (Figure 6). Pathogens causing NTDs accounted for 5% of the species tested. Only two studies focused on Leishmania spp. specifically. The first used fluorometric LAMP to detect 0.1 pg of L. donovani DNA (Puri et al. 2021). The other study used electrochemical sensing coupled to RPA, detecting 10 copies/µl (<1 pg) of L. braziliensis DNA (Wu et al. 2023). Therefore, more work needs to be done testing these new devices in the diagnosis of CL and other NTDs where they are desperately needed. Artificial intelligence More recently, machine learning has also been tested for visual classification of CL lesions. Promisingly, studies in Brazil, Cameroon, and Colombia showed 93–96% accuracy, although a study in Libya showed only 70% accuracy (Arce-Lopera et al. 2021; Steyve et al. 2022; Mohamed Noureldeen et al. 2023; Leal et al. 2024). Whether this is due to species-specific differences in lesion appearance, differences in skin tone, or other discrepancies in the experimental approach is currently unknown. Notably, only the Brazilian study took the causative species used in the training set into account, which the authors identified as L. braziliensis, L. amazonensis and L. guyanensis (Leal et al. 2024). Therefore, more work remains to make sure that these approaches have high accuracy in a range of different endemic settings. In a different approach, a study in Iran tested three different algorithms and found that they could distinguish CL lesions which were responsive or non-responsive to treatment with 56–77% accuracy (Bamorovat et al. 2025). AI has also been used to visually diagnose other cutaneous NTDs – Buruli ulcer, leprosy, scabies, mycetoma and yaws (Yotsu et al. 2023). Buruli ulcer and scabies were correctly diagnosed with 88% sensitivity and yaws was correctly diagnosed with 79% sensitivity. However, specificity was very poor. Mycetoma was misdiagnosed as Buruli ulcer in 41% of cases, leprosy was misdiagnosed as scabies in 64% of cases and scabies was misdiagnosed as yaws in 14% of cases. Leprosy and mycetoma were correctly diagnosed in only 13% and 0% of cases, respectively. AI has also been deployed in the detection of Leishmania spp. parasites by microscopy, with 73% to 99% accuracy (Abdelmula et al. 2024; Contreras-Ramírez et al. 2024; Kumar et al. 2024; Sadeghi et al. 2024; Tekle et al. 2024; Gadri et al. 2025). This represents a rapid improvement as a study from 2022 found 60% accuracy for infected macrophages (Zare et al. 2022). Although this approach promises to substantially reduce workload and minimise human error, it does not eliminate many of the previously discussed shortcomings of microscopy-based diagnostic techniques. In fact, only one of the studies cited used a smartphone camera to capture the images from the microscope, while the others used high resolution digital cameras (Gadri et al. 2025). Promisingly, the use of a smartphone camera did not compromise accuracy, which was 97%. Nevertheless, other issues such as the need for microscopy facilities and technical expertise in preparing the sample limit the applicability and relevance of these developments in a point-of-care setting. Telemedicine The use of mobile apps and telemedicine to augment CL diagnosis is another emerging field. A recent review on the use of telemedicine in NTDs found that it increases the capacity of the healthcare system to track the disease (Salvador et al. 2025b). One of the case studies was an app developed in Colombia for the use of patients and healthcare professionals. The app made it easier to follow-up on patients’ disease progression, allowed the collection of data in the form of lesion photographs and reporting of adverse events and made patients more likely to adhere to treatment (Cossio et al. 2023). Another study in Brazil found a high rate of patient satisfaction in using telemedicine for a CL consultation, with one of the advantages listed as saving money on travel costs (Salvador et al. 2025a). The impact of the WHO’s own SkinNTDs app was recently studied in Cameroon (Moungui et al. 2025). The app scored highly on user satisfaction (3.3/4.5), quality of information (3.6/5) and impact on knowledge and diagnosis (3.9/5), although 57% of participants had limited experience with skin NTDs and reported encountering less than one case per month. In a comparable study conducted in Ghana and Kenya, where only 22% of participants reported encountering less than one skin NTD case per month, the app scored 3.8/5 on user satisfaction, 4/5 on quality of information and 4.5/5 on impact (Frej et al. 2022). While these reports are encouraging with regard to skin NTDs generally, it would be useful to see equivalent data in countries which are more affected by CL in particular. Another useful tool is Lesionia, an example of Digital Systems for Data Management, which allows clinicians and healthcare workers in different countries to log and record epidemiological data on CL (Harigua-Souiai et al. 2025). The AI tools discussed above can be integrated with a telemedicine app to allow for a rapid triage step which can then lead to a referral for further testing. The use of AI image analysis tools in this way is already recommended by the WHO in the diagnosis of tuberculosis (World Health Organization, 2022). Remote diagnosis of CL by a mobile app was also successfully used on a small scale in Nepal during the COVID-19 pandemic (Parajuli and Prajapati, 2023). Conclusions A number of advances have been made in recent years that bring the goals of the TPP closer to being realised. Most of the ingredients for a good diagnostic test for CL already exist, albeit scattered across different disciplines and research areas. The serological CL Detect RDT has achieved 93% specificity over various studies, but its sensitivity of 68% is far below the 95% target in the TPP (Gebremeskele et al. 2023). The isothermal Loopamp RDT improved on this substantially, with 93% sensitivity and 87% specificity (Taye et al. 2025). This is also an improvement on qPCR which was found to have 87% sensitivity and 87% specificity (Rihs et al. 2025a). However, both RDTs performed very poorly on samples from Ethiopia, which highlights the country as being particularly neglected in terms of readily available diagnostic tools (Van Henten et al. 2022; Taye et al. 2024). The performance of the Loopamp test on crude VL samples indicates that a similar approach could be applied to CL samples, which would simplify sample preparation as outlined in the TPP. Additionally, a number of isothermal amplification devices matching the $2000 price target have now been developed. These could easily be integrated with a telemedicine app that uses AI-based visual lesion analysis as a triage step. This could have a dual effect of improving specificity by ruling out other lesion-causing dermatoses, but also meeting the goal of real-time connectivity, also listed in the TPP. Finally, the test result and lesion photograph could be logged together into a global database, to be regularly updated with patient information about the duration of the disease and responsiveness to treatments. In order to complete the process within the timeframe of the 2030 Roadmap, a focused and concerted effort remains to bring all of these elements together into a well-validated and coherent diagnostic algorithm. Supporting information Jarzabek and Denny supplementary material Jarzabek and Denny supplementary material