Bone Marrow Aspirates Alone without a Trephine Biopsy May be Insufficient for the Detection of Residual Leukemia Following Intensive Chemotherapy for the Treatment of Acute Myeloid Leukemia

Background: The diagnosis of acute myeloid leukemia (AML), response to treatment and disease recurrence are most commonly assessed with bone marrow studies. Recommendations from leading experts (Bain, 2001) and guidelines of the European LeukemiaNET (Dohner, 2010) and the National Comprehensive Cancer network (O’Donnell, 2012) suggest that only the bone marrow aspirate (BMA) is necessary to assess residual disease while the trephine biopsy (TB) is necessary only when an aparticulate BMA is obtained. In contrast, guidelines of the International Council for Standardization in Hematology (Lee, 2008) suggest that BMA and the TB should be routinely performed together as they provide complementary information. Due to these conflicting recommendations we sought to determine whether the TB provides additional sensitivity for the detection of residual leukemia following intensive chemotherapy for AML.

Methods: A single centre retrospective chart review was conducted of bone marrow studies of all AML patients who had received intensive chemotherapy from 2004 – 2013. Those lacking a TB were excluded. Residual disease was assessed by morphological examination of the BMA and TB +/- immunostaining but minimal residual disease (MRD) analysis was not performed.

Results: 598 bone marrow studies from 227 patients were evaluated. The median age of the patients was 54.6 (range 18 -77) with 70% age < 60. Forty-four percent were female. Cytogenetics were favorable in 30 (13%), intermediate in 146 (64%), high-risk in 44 (19%) and failed in 7 (3%) of the patients. Of the 598 bone marrow samples 198 (33%) were interim marrows performed 14 days following initiation of induction or re-induction chemotherapy (D14 marrow), 251(42%) were recovery marrows following induction/re-induction chemotherapy (EOI marrow) and 149 (25%) were during follow-up.

The BMA was considered to be acellular/hypocellular in 31%, hemodilute in 16.4% and aparticulate/pauciparticulate in 27.3% of samples. As per guidelines > 200 cells were counted in 99.8% of the aspirate samples to ascertain remission status. The median length of the TB segments was 1.85 cm (0.2 – 7.0 cm) and it was considered inadequate in 12.7%, of good or excellent quality in 24.9% and adequate for residual disease assessment in the remainder of cases. Approximately 19 % of TB samples had mild to significant hemorrhagic artifact. The bone marrow cellularity could not be assessed in 1.2% of samples but was patchy in 0.5%, aplastic in 2.8%, hypocellular in 36%, normocellular in 23.6%, hypercellular in 23.2%, packed in 6% and was not described in 6.7% of the cases.

Residual leukemia was identified in 33.1% of BMA and in 33.3% of TB samples. The BMA and the TB findings were concordant in 562 of 598 (94%) of cases. In 3.5% (21) of cases residual leukemia was seen in the TB but not the BMA whereas in 2.5% (15) of cases the BMA detected residual disease but the TB failed to do so.

The TB led to a change in diagnosis from ‘No Leukemia’ to ‘Residual Disease’ in 5.1% of D14 marrows, 3.6% of EOI marrows and in 1.3% of follow-up marrows with no statistically significant difference between the groups (p=0.178). There was no relationship between a change in diagnosis and whether patients received an anthracycline or a non-anthracycline based chemotherapy regimen (4.4% vs. 3.2%, p=1.0). The TB, however, led to a change in diagnosis more commonly in patients with favorable risk karyotype relative to those with intermediate risk karyotype (20% vs. 6.2%, p= 0.02) but not relative to those with unfavorable risk karyotype (20% vs. 13.6%, p=0.53).

Hemodilute bone marrow samples were more likely to have a TB related change in diagnosis relative to undilute samples (8.2% versus 2.6%, p=0.01) as were aparticulate/pauciparticulate samples relative to particulate samples (8% vs. 1.9%, p=0.00046). However, in multivariate analysis, only an aparticulate/pauciparticulate sample was associated with TB related change in diagnosis (p=0.015, OR = 3.6).

Conclusions: Our data demonstrate that, following intensive chemotherapy, the BMA alone may fail to identify residual leukemia particularly when the BMA is aparticulate/pauciparticulate. In these situations the TB provides additional sensitivity for the detection of residual disease. Further studies are required to evaluate the need for the TB in particulate samples when combined with MRD analysis.