DeepCxR would cut TB screening costs at some loss of accuracy

Assessors for India's HTAIn estimated that DeepCxR saves 51,330 rupees per 1,000 people screened for TB against radiologists, with specificity of 62.6% against 89%.

  • Not rated an organization

    The saving is a modeled estimate per 1,000 people screened, and the brief reports no saving realized in a screening program.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Partly reported benefit.
Evidence

foundstudiedreportedclaimed

Only the evaluators support the benefit.

Compared with

Radiologists' interpretation of digital chest X-rays, modeled for 1,000 adults with presumptive TB. Assumed without a measurement.

Without the system
Sensitivity 87%, specificity 89%, cost 299,077 INR per 1,000 screened
With it
85.9% Sensitivity and 62.6% Specificity, cost 247,747 INR per 1,000 screened
The system’s part
The model sets screening with DeepCxR against radiologists reading digital X-rays, which isolates the tool's part, though only in a model. The brief does not attribute its accuracy inputs to a named study. It describes AI-driven pattern recognition, and neither the brief nor ICMR's portal says that the tool learns from data.
Not yet checked
LimitsCosts and harmsLearns from dataLater outcome

Download data Cite Suggest a correction

Benefit type
Reduced Task Cost
System
DeepCxR (AI-assisted chest X-ray interpretation for TB screening)
Made by
Indian Council of Medical Research (ICMR), Department of Health Research (DHR) and Institute for Plasma Research (IPR), Department of Atomic Energy, Government of India
Used by
India's National TB Elimination Program (NTEP), as the program that HTAIn modeled (no deployment described)
Where
India
Stage
Tested in a simulation
Date
, when the source was published
Kind
Aggregate result
Broad benefit
Identifying Objects, Actions, and Events
Collections
Scrutiny and findings
Last revised
1 November 2026

Sources

  1. Assessment report, by an evaluator that the makers or users commissioned.

    Indian Institute of Public Health Gandhinagar, Resource Hub for Health Technology Assessment in India (HTAIn), Department of Health Research, Ministry of Health and Family Welfare. Health Technology Assessment of AI-Assisted (DeepCXR) Chest X-Ray Interpretation for Pulmonary Tuberculosis Screening. Policy brief, August 2026. Conclusion, p. 2. DHR, which runs HTAIn, co-developed DeepCxR.

    DeepCxR is less effective (85.9% Sensitivity and 62.6% Specificity), but also less costly (51,330 INR cost- saving per 1000 individuals screened) than standard radiologist interpretation.

  2. Other, by the makers. Context or contrary evidence.

    Indian Council of Medical Research. About AI4CXR: Artificial Intelligence tool for Chest X-ray Screening, ICMR AI4TB X-ray Portal. Read on 1 November 2026.

    The Indian Council of Medical Research (ICMR), Department of Health Research (DHR), and the Institute for Plasma Research (IPR), Department of Atomic Energy, Government of India, have jointly developed an artificial intelligence tool, DeepCxR version 1.1, to be used as a chest X-ray screening tool.

Details

AI system

Workflow

DeepCxR read digital chest X-rays of adults with presumptive TB and flagged TB-related abnormalities. The model compared the costs and correct classifications of DeepCxR and of radiologists for 1,000 people screened.

Money

Kind of value

Modeled cost saving per 1,000 individuals screened for TB, against radiologist interpretation

Amount

Estimate: 51,330 INR

Recipients and payers

India's TB screening services

Follow-up

Conditions for replication

People screened are adults with presumptive TB, with cough or fever for more than 2 weeks, weight loss or coughing up blood. Confirmatory tests such as nucleic acid amplification tests follow a positive reading. Health services accept more false positives in exchange for the saving.

Entry details

Sector

Healthcare.

Purpose

ICMR, DHR and the Institute for Plasma Research built DeepCxR to sort chest X-rays into normal and abnormal and to flag findings such as TB in screening of high-risk groups.

37 more questions not yet checked

model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, currency and prices, period and unit, realization, gross or net, included costs, comparison, method, return and payback, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). DeepCxR would cut TB screening costs at some loss of accuracy (AIBD-1164). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-1164.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)

BibTeX Data Field definitions