Authors' objectives:

The Australian Health Technology Advisory Committee completed a review of nucleic acid amplification (NAA) technology in 1997. The use of polymerase chain reaction (PCR) was not recommended for diagnosis and monitoring of acute myeloid leukemia (AML) in that report. MSAC was approached by the Pathology Services Table Committee (PSTC) to review the use of NAA in AML as well as the other indications considered in the 1997 AHTAC report. The NAA supporting committee developed the following questions for AML:

- Does AML1-ETO and CBF beta-MYH11 detection increase the proportion of patients who are recognised to have a specific disease entity that defines a specific therapeutic strategy?

- Does repeated qualitative or quantitative PCR testing in AML predict relapse and/or influence therapeutic decision making?

Authors' results and conclusions: Diagnostic accuracy in AML diagnosis PCR was evaluated for its ability to correctly identify patients with AML1-ETO and CBF beta-MYH11 AML at presentation. The 25 studies identified and included were all case series (level IV evidence). The sensitivity of cytogenetic and PCR testing was high for both abnormalities. PCR testing detected extra cases not identified by cytogenetic testing. For AML1-ETO 3.0 per cent (95% CI 1.9, 4.6) of all AML cases were PCR positive and cytogenetic test negative. Likewise, for CBF beta-MYH11 1.4 per cent (95% CI 0.9, 2.0) of all AML cases were PCR positive and cytogenetic test negative. All tests (including cytogenetic testing) are considered to be imperfect reference tests due to their imperfect sensitivity, so the finding of PCR positive and cytogenetic negative cases is not unexpected. For instance, all alternative tests require the presence of a higher proportion of abnormal cells to detect AML1-ETO or CBF beta-MYH11 than is required by PCR. Therefore, the discrepant proportion probably represents the incremental benefit of adding PCR testing to cytogenetic testing, although the estimates presented may over estimate this benefit due to the low quality of the studies included. PCR was estimated to have a specificity of 99.6 per cent (95% CI 97.8, 100) for AML1-ETO and 99.2 per cent (95% CI 97.3, 99.9) for CBF beta-MYH11. Diagnostic accuracy in AML monitoring PCR was evaluated for its ability to predict subsequent cytogenetic and haematological relapse in the monitoring studies. Nine studies were eligible for the review of monitoring for AML1-ETO and seven for CBF beta-MYH11. All were case series (level IV evidence). Six of the AML1-ETO articles and three of the CBF beta-MYH11 articles used quantitative PCR methods. The pooled diagnostic odds ratio (DOR) for AML1-ETO was 8.7 (95% CI 3.4, 22.0) and for CBF beta-MYH11 the DOR was 35 (95% CI 10, 119). A DOR of 8.7 is consistent with, for example, a sensitivity of 80 per cent and specificity of 69 per cent while a DOR of 35 is consistent with a sensitivity of 95 per cent and specificity of 65 per cent.

Authors' recommendations: MSAC recommended that on the strength of evidence pertaining to the safety, effectiveness and cost-effectiveness of polymerase chain reaction (PCR) testing in the diagnosis and monitoring of acute myeloid leukaemia (AML), public funding should be supported for this procedure.

Authors' methods: Systematic review

Project Status: Completed

URL for project: http://www.msac.gov.au/

Year Published: 2003

English language abstract: An English language summary is available

Publication Type: Not Assigned

Country: Australia

Organisation Name: Medical Services Advisory Committee

Contact Address: MSAC (MDP 107), GPO Box 9848, Canberra, ACT 2601, Australia. Tel: +61 2 6289 6811; Fax: +61 2 6289 8799.

Contact Name: msac.secretariat@health.gov.au

Contact Email: msac.secretariat@health.gov.au

Copyright: Medicare Services Advisory Committee (MSAC)