You are here

BCR/ABL ratios % and IS factors


Hi ,

I have read the primer which is on this site and it is quite useful but nevertheless would like some comments from those of you that might have more expertise/experience than myself.

I was dx more than 14 years ago and have access to my BCR/ABL scores for the whole period which on perusal makes interesting reading from a 105% ratio from the first bone marrow to some more recent undetectables and with loss of MMR on two occasions.I have had a mixture of some private treatments and some on NHS ,always on imatinib BUT always using the same laboratory-Hammersmith;however I have not been treated there as a patient.Up until April 2019 the IS factor quoted for that lab has been 0.299 but from that date the I.S.Factor has changed to 0.405 -a note on my latest BCR/ABL report states "Due to a change in equipment ,this test is awaiting UKAS assessment".So it looks as if due to replacement and updating of equipment some new calibration has taken place but is subject to further review and assessment.

Now if the initial ratio or % is very small or nearly undetectable then multiplying it by an I .S. factor changes it to some extent

Example: BCR/ABL ratio 0.1 x I S Factor 0.299 = Ratio I.S. 0.0299  My observation would be 0.1 is safe territory but when converted to an IS ratio it looks even safer.

In one instance my BCR/ABL was 0.4 indicating a loss of MMR but 0.4 x 0.299= 0.1196; the raw ratio would be a cause for concern but the adjusted one perhaps would not be.My specialist at that time considered moving me on to another tki such as nilotinib but considering the Ratio I.S.was just above 0.1  did not. A subsequent specialist became quite concerned and ordered mutations analysis ( none found) and upon advice taken from Hammersmith put me on high dose imatinib of 600 mg which then eventually led to some undetectables. So is there still some doubt over the accuracy of I.S.Factors?

I am still unsure whether one should still look very carefully at the raw BCR/ABL ratio or very much depend on the the adjusted ratio that takes into account the I.S. Factor.I fully understand the nature of the PCR test being very sensitive and also the need to try to develop a level playing field between labs ;in addition if one switches labs then one needs to have a comparison tool.However if one has been with a single lab for 14 years I would have questions about the need for/value of  an IS factor and also reasons to change it -as seems to have happened recently at Hammersmith.

Any comments or advice from those who may have had similar concerns?

Best wishes



Hi John, having managed a UKAS accredited microbiology lab, I hope I can help. As part of the UK Accreditation Service audit (against the standard ISO 17025), equipment has to be shown to be operating within defined levels of accuracy. New equipment will often need some time to accumulate the necessary data to complete the statistical analysis. If you have been having tests for 14 years, this is likely at least the second change of equipment the lab has had in that time. At each change the kit will have a different IS conversion factor, so although the tests are all done at the same lab, the equipment change effectively makes it a new lab for comparison purposes.

A good while ago (7-8 years) there were a few people on here who had been undetectable for years, but started getting low positive results. That for me tied up with labs getting upgraded equipment which was more sensitive, and could therefore detect at levels that old equipment would not detect. It happened to me; I knew that one the manufacturers had new kit out and got my consultant to ask the lab the question.

It is good that you know the kit has changed and a new factor will come out.

UKAS accredited labs have to participate in what the standard calls proficiency trials; normal people would call them round robin testing exercises. A sample is prepared, and divided between the participating labs. All the results are analysed by clever statisticians, for whom I have the utmost respect.  They can show any labs which have results outside the range of all the others, and give the data for cross calibration between labs. 

Thank you Alastair for this explanation.

Hi Alistair,

Thank you for your expert input .One of the benefits of new equipment it seems is a faster turnaround time for the results to emerge .When long ago I was dx I was advised that it took 2-3 weeks for PCR  results to come back which if one was on a 3 monthly frequency of the usual bloods (full blood count,liver bone etc) and face to face with specialist the PCR was historical from over 2 months ago.I note now the turnaround time is for my lab is now about one week .

There still seem to be the issues of spoilt samples not sent in on time to the specialist lab and /or inadequate size of sample taken or not stored in the purple phial;my current specialist suggested to me that where patients use their own GP to take bloods he still has (pre Covid ) some issues of procedures not being followed by the phlebotomist.

Hopefully with updated machines/equipment there is greater  sensitivity and accuracy of results compared to before.There are some interesting presentations on You Tube of the process involved in the PCR

With best wishes,


Hi John, glad the input was useful. Certainly process times have gone down over the years; I was involved around 15 years ago in the first implementation of a same day forensic DNA profiling system. That is now down to around 90 minutes I believe. Sample size requirements have also reduced, but as you say sample integrity post collection can still be an issue. When I was on monthly BCR ABL reducing my dose the general phlebotomy service at the district hospital only got 50% of the samples right - I had to go the phlebotomist on the haematology  clinic to make sure the samples were collected and processed correctly.

In the PCR game the accuracy of the process is, for geeks like me, divided into a number of components which include precision (how close is the result to the right answer), repeatability (If you run the same sample again on the same kit with the same people how close do you get the same answer) consistency (if the same sample is run by other people on different kit how close do you get the same answer) and sensitivity (how many decimal places can the process detect and measure to reliably). I think the test is now as sensitive and precise as we need for CML patients. The fact we need IS conversions says that there is mileage in trying to improve repeatability and consistency.