The Impact of a Reflex Testing Strategy for High Risk Chromosome 14 Abnormalities on FISH Laboratory Costs in Multiple Myeloma

Background: The detection of cytogenetic aberrations by fluorescence in situ hybridization (FISH) has prognostic value in multiple myeloma (MM), with the presence of del(17p), t(4;14) or t(14;16) currently defining high-risk chromosomal abnormalities (HR-CA) in the Revised International Staging System. Both t(4;14) and t(14;16) involve the IGH gene, and can be detected by FISH probes targeting IGH-rearrangements; these probes, in turn, can also detect other non-HR-CA involving this gene.

Additionally, the interpretation of FISH probing is observer-dependent and influenced by the type of probe used. While dual-color dual-fusion (DC/DF) probes are more sensitive, break-apart (BA) probes are easier to interpret.

Aims: We aim to analyze the economic outcome of switching from a DC/DF detection of HR-CA of chromosome 14 (Cr14) to a strategy of a BA detection of IGH rearrangements with reflex testing for t(4;14) and t(14;16) only in positive samples.

Methods: We analyzed all FISH panels performed in our Lab for MM between June 1^st 2015 and July 31^st 2019. Until May 2016 we performed an a priori MM FISH panel consisting of probes for t(4;14), for del(17p), and for 1p/1q amplifications and deletions (Cohort 1); testing for t(14;16) was performed under request. From May 2016 to October 2017 we added a probe for t(14;16) to the routine panel (Cohort 2); during both periods, probing for IGH rearrangements was performed reflexively in the case of an inconclusive or otherwise uncertain t(4;14) and/or t(14;16) result. From November 2017 onwards we have probed for IGH rearrangements a priori, with reflex testing for t(4;14) and t(14;16) only in positive cases (IGH Cohort); testing for del(17p) and 1p/1q was unchanged throughout the timeframe and is not further analyzed in this study.

We looked at the costs associated with the acquisition of the relevant FISH probes, as well as the number of hybridizations performed per patient, for each of the three strategies.

Results: Over the time period under analysis we performed FISH panels on 450 MM patients (55% male); 21% of samples were from Cohort 1, 45% were from Cohort 2 and 34% were from the IGH Cohort.

Only 28.9% of patients with a documented IGH rearrangement were positive for either t(4;14) or t(14;16).

The mean cost per patient for IGH probes was 3.8±11.6 € in Cohort 1, 1.3±7.1 in Cohort 2 and 36.2±10.1 in the IGH Cohort (p<0.001). The mean cost for t(4;14) probes was 44.1±0, 44.1±0 and 22.0±22.1 €, respectively (p<0.001), and the mean cost for t(14;16) probes was 3.2±9.9, 31.2±8.5 and 16.3±16.8 €, respectively (p<0.001). Overall, the mean cost per patient for all three probes was 51.1±18.7, 76.6±9.5 and 74.5±33.3 €, respectively (p<0.001 for the pre- and post-t(14;16) comparison; and p=NS for the pre- and post-IGH comparison).

The mean number of hybridizations performed per patient was 1.0±0.2 prior to the introduction of a priori IGH testing, and 1.4±0.5 in the IGH Cohort (p<0.001).

Discussion: We found that the addition of routine testing for t(14;16) increased the mean cost of probes per patient by nearly 50% (from approximately 50€ to 75€ - US$55-85), while the switch to a priori testing for IGH rearrangements with reflex probing for the translocations did not increase costs with FISH probes, as the overall higher number of tests performed was balanced by the lower cost of the BA probe compared to the DC/DF probes, as well as a reduction in DC/DF probe consumption.

On the other hand, after the switch, the number of hybridizations performed per patient increased 40%, corresponding to an increase in 40% in both laboratory technician and pathologist man-hours, and ancillary (non-probe) reagents spent.

Less than two-thirds of patients with an IGH-rearrangement were positive for one of the two HR-CA of Cr14. This reflex testing strategy, therefore, has the clinical benefit of identifying patients for probing for additional Cr14 translocations of putative prognostic value, such as t(11;14).

Conclusions: We found that this reflex testing strategy was associated with a 40% increase in non-probe reagents and in technician man-hours, but no significant impact in probe costs. This must be weighed against the potential clinical and scientific gains from the detection of additional Cr14 aberrations with thus-far uncertain or untested prognostic value.