Addition of CT Scans Fails to Prove Predictive Power of CR,PR

[Guest guest]

[2833] The Addition of CT Scans to NCI-96 Chronic Lymphocytic

Leukemia (CLL) Response Criteria Fails To Improve the Predictive

Power of Complete Response (CR) and Partial Response (PR) as Assessed

by Improvement in Progression-Free (PFS) and Overall Survival (OS) .

Session Type: Poster Session, Board #62-III

Kristie A. Blum, Donn Young, Broering, Margaret S. Lucas,

S. Lin, R. Grever, C. Byrd Division of

Hematology/Oncology, The Ohio State University, Columbus, OH, USA

Background: NCI-sponsored working group (NCI-WG) criteria for

response assessment in CLL (Blood 1996:87; 4990-4997) rely on

physical examination, blood, and bone marrow evaluations, with no

requirement for CT scan measurements. Recent improvements in CLL

therapy provide justification for re-evaluation of these response

criteria.

Methods: In a retrospective review of 3 ongoing trials at The Ohio

State University in patients (pts) with symptomatic untreated (8) and

previously treated (74) CLL, we assessed response in pre- and post-

treatment CT scans using international non-Hodgkins lymphoma (NHL) CT

response criteria (JCO 1999:17;1244-1253). These results were

compared to response determined by conventional NCI-WG criteria.

Results: In 82 pts, the median age was 60 (range, 38-84), with 60

males. The majority of pts had Rai stage IV disease (39), with a

median of 4 prior therapies (range 0-14). Response according to NCI-

WG criteria were as follows: CR 5 (6.1%), PR 32 (39%), SD 21 (25.6%),

PD 17 (20.7%), and not evaluable (NE) 7 (8.5%). Response by NHL CT

criteria were as follows: CR 3 (3.7%), CRu 12 (14.6%), PR 16 (19.5%),

SD 26 (31.7%), PD 4 (4.9%), and NE 21 (25.6%). 23 pts were NE by

either NCI-WG or NHL CT criteria due to lack of a post-treatment CT

or early treatment-related toxicity. The table below provides the

discrepancies in response assessment between NCI-WG and NHL CT

criteria in the 59 evaluable pts.

In all 82 pts, the median OS and PFS were 25.2 mos and 9.0 mos,

respectively. In the 75 pts evaluable by NCI-WG criteria, CR, PR, SD,

and PD response designations correlated with PFS (p=0.001) by log-

rank test, with a median PFS of 27.3 mos for CR, 11.4 mos for PR, 5.6

mos for SD, and 2.5 mos for PD. Response by NCI-WG criteria also

correlated with OS (p=0.012), with 12 deaths reported.

For the 61 pts evaluable by NHL CT criteria, log-rank test

demonstrated a significant correlation with CR, PR, SD, and PD

designations and PFS (p=0.001); however, there was no difference

between CRu and PR. For pts achieving a CRu, PFS was 11.7 mos;

compared to 14.7 mos for a PR, and 18.4 mos for a CR. NHL CT criteria

also correlated with OS (p=0.001), with 6 deaths reported. In multi-

variate cox regression analysis, NCI-WG CLL response, NHL CT

response, age, Rai stage, and number of prior therapies failed to

correlate with OS (p-values=0.382-0.999). However, with the same

variables, NCI-WG CLL response was the only significant predictor of

PFS (p=0.005), with no significant association between NHL CT

response and PFS (p=0.061).

Conclusions: These results support current NCI-WG CLL criteria as a

significant predictor of response and ultimately PFS in pts with

previously treated CLL, with no additional benefit from the inclusion

of CT scans. In addition, the CRu response definition has limited

prognostic value in CLL, as PFS in this group is similar to those pts

achieving PR. Limitations of this study include retrospective design,

low CR rate, and a high proportion of previously treated pts. These

results nonetheless highlight the importance of prospective

validation of CT scans before routine inclusion in response criteria

for CLL.