Abstract

Abstract 2055

Background:

Although bone marrow aspiration and biopsy have been utilized in the diagnosis and staging of hematologic malignancies for almost a century, little is known about the potential impact of operator techniques on the quality of specimens obtained. We aimed to characterize the impact of operator techniques used for bone marrow biopsies, and to determine if any of these techniques are associated with specimen quality.

Methods:

From April to July 2012, we recorded operator data corresponding to all consecutive bone marrow biopsy procedures resulting in a specimen performed as part of routine care at the Dana-Farber Cancer Institute (Boston, MA). These data included type of needle used, whether or not the operator redirected the needle between aspirate and biopsy, patient position (prone or lateral decubitus), use of a drill, use of a measuring stylet, and volume of aspirate obtained in the first pull. We then reviewed the resulting pathology reports, focusing on specific indicators of quality such as presence of a diagnosis, spicularity, and core length. For the purposes of this study, we allowed credit for a diagnostic specimen if the pathology reports mentioned normal trilineage hematopoiesis or a pathologic process on either the aspirate or core biopsy. Univariate associations between operator techniques and bone marrow biopsy quality were determined using chi-square and Fisher's exact tests; multivariable logistic regression taking patient characteristics (age, gender and body mass index) and individual operator into account were also fit to characterize operator techniques that were independently predictive of quality. Finally, we assessed the association between spicularity and core length with diagnosticity.

Results:

413 procedures performed by a total of 23 operators were analyzed. 91.5% of the bone marrow evaluations were diagnostic, 66.6% were spicular, and 52.8% had a core length greater than 1cm. The univariate analysis of operator techniques is detailed below:

Operator Technique Diagnostic (%) Spicular (%) Core length >1cm (%) 
Rosenthal needle 86.0 p = 0.009 64.4 p = 0.51 46.3 p = 0.11 
Non-Rosenthal needle 93.8 67.8 54.9 
Redirect, Yes 90.4 p = 0.38 62.8 p = 0.04 46.9 p = 0.01 
No 92.8 72.4 59.6 
Prone position 91.5 p = 1.00 68.1 p = 0.02 53.2 p = 0.24 
Lateral decubitus 91.7 33.3 33.3 
Drill, Yes 92.3 p = 1.00 84.6 p = 0.23 61.5 p = 0.58 
No 91.5 66.2 52.1 
Stylet, Yes 91.6 p = 1.00 66.9 p = 1.00 52.2 p = 1.00 
No 100.0 80.0 60.0 
Aspirate volume, ≤1cc 92.0 p = 0.62 66.4 p = 0.68 50.2 p = 0.31 
>1cc 90.6 68.3 55.2 
Procedures, <9 per day 92.0 p = 0.74 66.2 p = 0.76 50.2 p = 0.37 
≥9 per day 91.1 67.6 54.7 
Operator Technique Diagnostic (%) Spicular (%) Core length >1cm (%) 
Rosenthal needle 86.0 p = 0.009 64.4 p = 0.51 46.3 p = 0.11 
Non-Rosenthal needle 93.8 67.8 54.9 
Redirect, Yes 90.4 p = 0.38 62.8 p = 0.04 46.9 p = 0.01 
No 92.8 72.4 59.6 
Prone position 91.5 p = 1.00 68.1 p = 0.02 53.2 p = 0.24 
Lateral decubitus 91.7 33.3 33.3 
Drill, Yes 92.3 p = 1.00 84.6 p = 0.23 61.5 p = 0.58 
No 91.5 66.2 52.1 
Stylet, Yes 91.6 p = 1.00 66.9 p = 1.00 52.2 p = 1.00 
No 100.0 80.0 60.0 
Aspirate volume, ≤1cc 92.0 p = 0.62 66.4 p = 0.68 50.2 p = 0.31 
>1cc 90.6 68.3 55.2 
Procedures, <9 per day 92.0 p = 0.74 66.2 p = 0.76 50.2 p = 0.37 
≥9 per day 91.1 67.6 54.7 

In multivariable models, no operator technique was significantly associated with obtaining diagnostic specimens; however, lateral decubitus position was associated with lower likelihood of obtaining spicular specimens (odds ratio (OR) 0.15 [95% confidence interval (CI) 0.04, 0.59], p=0.007), and redirecting was associated with lower likelihood of core length >1cm (OR 0.41 [95% CI 0.20, 0.82], p=0.01). Spicular specimens had significantly higher rates of diagnosticity compared to aspicular specimens (96.3% vs. 71%, p <0.001). Finally, among the thirty-five patients who had non-diagnostic biopsies, nine underwent repeat biopsy within thirty days of initial procedure.

Conclusion:

In our large cohort of observed bone marrow procedures, we found a high diagnostic rate of 91.5%. Moreover, although there were several univariate associations, in multivariable models, specific operator techniques did not predict higher likelihood of obtaining a diagnostic specimen. Although redirecting was associated with shorter cores, core length was not significantly associated with diagnosticity. Overall, our data suggest that several different operator techniques ultimately result in bone marrow specimens of similar quality.

Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.