CAD for TB: proven
Artificial Intelligence

Artificial intelligence is increasingly applied to medical image analysis. CAD4TB is the first algorithm for chest radiography analysis that is validated in various operational research publications, both clinical and technical. Computer Aided Detection for TB developed by Dutch Thirona / Radboud University outperforms trained human readers to detect and quantify the probability of the presence of abnormalities consistent with TB [1]. CAD can be used as a 1 minute low cost triage tool before Xpert for individuals from 4 years of age.

Partial financing for this R&D was obtained from the Dutch Ministry of Economic Affairs & Innovations with support from CheckTB! CAD4TB can be used offline on a laptop on site or in the cloud. CAD4TB is the only scoring software that has been evaluated and is being implemented in programmatic settings. Encouraging findings on the diagnostic accuracy of CAD4TB has been reported from sub-Saharan Africa, and most recently from Bangladesh [2]. Also other suppliers of computer assisted reading software for chest radiographs are now available, but these can be considered to be in the earlier phase of operational, technical and clinical evidence build-up.

At the recent UNION Conferences CAD4TB sensitivity and specificity scores on data sets from Bangladesh, Cameroon (Stop TB Partnership/TB REACH), Nambia (TB prevalence survey), Philippines (WHO), Tanzania (USAID), South Africa (GLOBAL FUND) and Zambia (TB prevalence survey) were presented, confirming the potential of CAD4TB as a rapid triage before Xpert MTB/RIF. CAD4TB is currently under evaluation by WHO as part of a systematic review. CAD designed to support automated processing of chest x-rays for early detection, monitoring and quantification of pulmonary TB, can in the future also be used to detect abnormalities suggestive of silicosis, pneumonia, emphysema and lung cancer.

Digital imaging innovations for early TB case detection was presented by CheckTB! at the UNION World Conference Kuala Lumpur for the New Diagnostics Work Group on November 13th, 2012.

CAD design

CAD4TB™ was developed to automatically detect tuberculosis-related abnormalities in posterior anterior chest X-rays. This computer-aided detection software takes a single chest X-ray as its input, in the form of a DICOM image, and produces several outputs: a quality assessment of the input image, a heat map highlighting possible abnormal areas, and a score between 0 and 100 indicating the likelihood of the X-ray being abnormal and the subject on the X-ray being affected by tuberculosis.

The CAD system follows the general machine learning approach of supervised learning. In supervised learning the system learns from a set of examples, called the training data. The training data consists of a number of labelled images of which the TB status is known. For the CAD4TB system there are two classes and the labels are normal and suspect for TB. A number of features is extracted from these images and combined to form a feature vector for each image. These feature vectors are used to train a TB classifier. This TB classifier incorporating state-of-the-art pattern recognition techniques can then be used to classify images of which the label is not known.

Computer Aided Diagnosis will enable automated accurate pre-screening on site by the digital X-ray system allowing radiologists or pulmonologists to focus more on analysing the TB suspect images. In addition CAD4TB can contribute to fewer "over-reading" by less experienced radiographers or clinical officers. CAD in combination with molecular tests such as Xpert innovates screening for early case detection. This CAD-Xpert combination enables rapid and lower cost active case finding in high risk groups such as PLWH, diabetics, mine workers and inmates.

Use of CAD4TB

To increase TB prevalence surveys' accuracy and to reduce the cost of risk group screening programs CAD offers the following diagnostic functionality:

  • quality assessment of the chest image
  • provide a probability score of abnormalities consistent with TB between 0 and 100
  • provide markers around suspect regions
  • integrate score and dCXR in a PACS and/or electronic patient file

CAD4TB also allows for pre-selecting its sensitivity, which is of special value for prevalence surveys. Now that the CAD4TB algorithm is optimized the number of CXR requiring human reading can be reduced to less than 20%. In TB prevalence surveys in Botswana and Mozambique CAD4TB will be used as the radiological gold standard.

The progress

The “Computer Aided Detection for TB (CAD4TB)” project is still in development as the detection of other lung diseases by the algorithm will further increase its specificity. The Radboud University of Nijmegen/Thirona is now involved on the Dutch side together with Delft Imaging Systems they collaborate to further optimize the accuracy of CAD using deep learning techniques.

In order to achieve this, tens of thousands of digital images taken with the same technology direct digital X-ray systems were analysed in the same systematic way. Various state-of-the-art pattern recognition techniques are applied to continuously improve sensitivity and specificity scores.

Expected Outcome

The objective of the joint research since 2002 was to achieve a CAD system with a sensitivity of 90% and a specificity of 80%. The specificity target was the biggest challenge for the scientists, but with CAD4TB v6 (2018, CE certified) this is overachieved on a large dataset of the Philippines. This will further enhance the role of digital X-ray as a screening tool to effectively detect presumptive TB.

CAD improves TB screening and supports active case finding by:

  • A high percentage of presumptive TB cases is identified automatically wherever the digital X-ray is used
    • Supporting radiographers and/or clinical officers with first image reading
    • Reducing number of human image interpretations and prevent under or over-reading
    • Providing a powerful offline tool in prevalence studies
    • Offering an effective quality assessment tool for health workers and NTP
    • Reducing screening and diagnostic costs
  • Only a limited number of complex Chest images may require human radiological expertise

While developed for TB, this system will be enhanced in the future to spot other lung diseases. It is thus expected to contribute to a broad healthcare strengthening impact in areas where this is most needed.

More than 30 countries already use CAD4TB in their TB programs. In below graph the role of CAD in these programs is illustrated [3]. IRD in Pakistan - one of the early adopters of CAD4TB with Stop TB Partnership/TB REACH support - is now planning to screen up to 5 million people using CAD4TB as a rapid triage also in support of the Zero TB initiative.

Namibia, Philippines, Southern Africa, UK and Tanzania

More recent studies with digital chest images from Namibia, Southern Africa, London Find & Treat, Zambia and Tanzanian TB prevalence surveys, confirm that CAD is already as accurate as the trained human reader. At the UNION Conference in Kuala Lumpur a linear relationship between CAD score and TB detection among different strata of patients was presented by Zambart confirming CAD as a screening tool before Xpert. On a large data-set of the Philippines CAD v5 scored 94 sensitivity at 84% specificity with Xpert MTB/RIF as reference. As presented at the 2018 UNION Conference, data from the Namibian TB prevalence survey confirmed that CAD4TB accuracy allows the trained human reader to be replaced by this AI algorithm in future surveys.

Study World Health Organisation in Philippines

A study was being performed on CAD4TB performance in the Philippines to automatically interpret digital chest images made for migrant TB screening programs. This research is a collaboration between Radboud University Medical Center and WHO Philippines with initial liaison through by KNCV Tuberculosis Foundation and CheckTB!.

12,256 CXRs from Palawan provincial areas of the Philippines [4]; CAD4TB version 4 achieved 90.0% sensitivity at 80.0% specificity*. In a publication CAD4TB version 5 will confirm its further increased accuracy, both sensitivity and specificity to augment to 94 and 84% respectively.

CAD4TB with Xpert® reference

*46th Union World Conference on Lung Health, 2015; Rick H.H.M. Philipsen et al Computerized Chest Radiography Screening to Detect Tuberculosis in the Philippines, 2015.

CAD4TB AI continuously being improved

As shown below the AUC of CAD4TB has significantly increased since the initial CAD version was released in 2013. The AUC gains are reported by Thirona/Radboud University for both radiological and bacteriological reference. This means that the accuracy of the AI algorithm significantly improved through gains in both sensitivity and specificity also as a result of applying deep learning techniques. As a key benefit for TB programs in operational use, CAD version 6 allows for the rapid detection of more (presumptive) TB cases at lower cost.

The ROC curves[5] below provide information on the pooled sensitivity and specificity scores of in total 4 consecutive CAD4TB versions again with reference radiological (expert human reader) and bacteriological (culture). As can be noted the performance in terms of sensitivity and specificity improves per each new version of the software. User can, also depending on the available budget for e.g. Xpert cartridges, with the CAD abnormality score threshold select the preferred trade-off between sensitivity and specificity.


From the various studies (please refer to CAD Publications section on this site) it can be concluded, that CAD4TB already performs similar or better than the human readers in classifying radiographs as being normal or suspect for TB. The performance of the CAD system is expected to further improve with more training. Version 6 of CAD4TB is already at expert c.q. radiologist level to classify abnormal chest images suggestive of TB. CAD as a rapid triage before Xpert has the potential to cost-effectively scale up systematic screening of high risk groups in support of the End TB Strategy. This supports the Stop TB Partnership Paradigm shift as also in low resource settings people at risk of TB can with AI support access high standards of TB care that were available for decades in Western Europe and North America.