Skip to main content
. 2025 Sep 3;16:1665685. doi: 10.3389/fmicb.2025.1665685

Table 2.

Comparison of conventional, molecular, genomic, and Artificial Intelligence (AI)-based diagnostic methods for diagnosis and differentiation of NTMs.

Conventional methods Molecular methods Genomic methods Artificial Intelligence (AI) based methods
  1. Radiography methods

    • CT or chest X-rays (Cao et al., 2021).

    • Routine screening or follow-up investigations

    • not cost-effective.

  2. Sputum cultures and AFB smear:

    • Sputum AFB smear and culture.

    • Drug susceptibility testing (DST) and genotypic identification depend on culture, the gold standard for laboratory validation of NTMs (Zaizen et al., 2022).

  3. Biopsy for AFB presence:

  4. Biochemical Tests:

    • Biochemical tests include reduction of niacin secretion, Arylsulfatase test, Nitrate reduction, Catalase activity, tween eighty hydrolysis, Citrate utilization, urease activity, reduction of tellurite for different NTM species (Bhalla et al., 2018).

    • Time-consuming and less-accurate hence has become obsolete and it has been replaced with sensitive molecular tests

  5. Blood culture with MGIT:

    • Culture in BD BACTEC™ MGIT™ is the technique used for disseminated NTM detection.

    • Growth detection is done by BD BACTEC™ MGIT™ 960 Instrument (Ferraro et al., 2024).

  1. Multiplex PCR:

    • Amplifies several target genes at once

    • Allows numerous NTM species detection in one experiment (Peixoto et al., 2020).

  2. Gene sequencing (rpoB, hsp65, 16S rRNA):

  3. PCR-RFLP:

    • Very useful in distinction of NTM species that have distinct fragment patterns.

    • It is quick, precise, and economical to identify NTMs (Nour-Neamatollahie et al., 2017).

  4. Repetitive element PCR (Rep-PCR):

    • Helps distinguish between closely related strains

    • Useful for strain typing, especially for non-mycobacteria that proliferate quickly, such as M. abscessus (Zhang et al., 2024a)

  5. Probe-based restriction analysis (PRA) or hsp65 PCR restriction enzyme analysis (PREA):

    • Frequently used to compare restriction fragment patterns to identify species within NTM group (Sajduda et al., 2012).

  6. MALDI-TOF MS:

    • It is a reliable test, but may have limitations in differentiating the closely related NTM species.

    • It might not be able to distinguish between closely related NTM (Alcaide et al., 2018).

  7. Line Probe Assays (LPA):

    • LPAs give faster results.

    • High sensitivity and specificity for identifying different NTMs

    • LPAs identify drug-resistant mutations in NTMs (Yoon et al., 2020).

  8. CRISPR-Cas12a:

    • Genome editing tool-based assays may be important in diagnosing NTMs

    • Cas proteins deliver nonspecific nucleotide cleavage to the intended location.

    • Produces a fluorescent signal that makes the target gene detectable (Chen et al., 2018).

  1. Whole-genome sequencing (WGS):

  2. Sequencing of specific genes:

  3. Metagenomic Next-Generation Sequencing (mNGS):

    • mNGS, identifies every nucleic acid in a sample

    • Allows for the identification of a wider variety of NTM species and mixed infection strains (Zhang et al., 2023).

  4. Targeted Next-Generation sequencing (tNGS):

    • Narrows down on specific genomic regions of interest leading to in-depth analysis of genetic variations.

    • Cost-efficient approach as well as rapid and comprehensive method (Buckwalter et al., 2023).

  1. Radiomics:

    • Extracts quantitative information for analysis from medical photos in a high-throughput manner.

    • A “radiomic signature” for NTM diagnosis is created by extracting quantitative information from medical images, such as CT scans (Al Bulushi et al., 2022).

  2. Machine Learning:

    • For developing prediction algorithms, machine learning can be a highly useful method.

    • Framework with an AUC of 0.9 that was based on a large number of lesions for radiomics tasks (Han et al., 2025).

    • A linear support vector machine (SVM) for machine learning tasks was used to evaluate the diagnostic significance of cavities and bronchiectasis in distinguishing NTMs from TB.

  3. Deep Learning:

    • Deep neural networks have been created to analyze chest X-rays and differentiate between NTM-PD and TB.

    • Finding insertions, deletions, and single nucleotide polymorphisms in the genome is made easier by deep learning algorithms.

    • Automatically extracts relevant information from complicated patient data, such as medical pictures (Liu et al., 2023).