- Email: [email protected]
Using Smudge Cells on Routine Blood Smears to Predict Clinical Outcome in Chronic Lymphocytic Leukemia: A Universally Available Prognostic Test
BRIEF REPORT
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
Using Smudge Cells on Routine Blood Smears to Predict Clinical Outcome in Chronic Lymphocytic Leukemia: A Universally Available Prognostic Test GRZEGORZ S. NOWAKOWSKI, MD; JAMES D. HOYER, MD; TAIT D. SHANAFELT, MD; SUSAN M. GEYER, PHD; BETSY R. LAPLANT, MS; TIMOTHY G. CALL, MD; DIANE F. JELINEK, PHD; CLIVE S. ZENT, MD; AND NEIL E. KAY, MD Recently developed prognostic tests in early Rai and Binet stage chronic lymphocytic leukemia (CLL) require considerable technologic expertise and are not available worldwide. Smudge cells are CLL cells ruptured during smear preparation. We hypothesized that smudge cell formation is inversely correlated with expression of vimentin, a cytoskeletal protein and prognostic marker, and that the percentage of smudge cells would predict prognosis in CLL. We reviewed the blood smears of 75 patients with previously untreated early- and intermediate-stage CLL (Rai stage 0-II) who were seen at the Mayo Clinic in Rochester, Minn, between September 1989 and December 2000. A total of 200 lymphocytes and smudge cells were counted on each slide and the results expressed as a percentage of the total lymphocytes (intact and smudged). The median percentage of smudge cells was 27% (range, 4%-72%). The percentage of smudge cells inversely correlated with vimentin expression (r=–0.57; P=.007). The median percentage of smudge cells was higher in patients with the mutated immunoglobulin heavy chain gene than in those with the unmutated immunoglobulin heavy chain gene (31% vs 13%; P=.02). Patients with less than 30% smudge cells had a median time from diagnosis to initial treatment of 72.7 months, whereas the median time from diagnosis to initial treatment in patients with 30% or more smudge cells was not reached (P=.001). The percentage of smudge cells as a continuous variable correlated with overall survival (P=.04). The estimation of smudge cells on a blood smear could be a universally available prognostic test in early-stage CLL.
Mayo Clin Proc. 2007;82(4):449-453 ALC = absolute lymphocyte count; CI = confidence interval; CLL = chronic lymphocytic leukemia; IgVH = immunoglobulin heavy chain; MFI = mean fluorescence intensity; TTT = time from diagnosis to initial treatment
W
ith the widespread use of automated blood cell counters and flow cytometric immunophenotyping, 60% to 80% of patients with chronic lymphocytic leukemia (CLL) now have their condition diagnosed at an early stage of disease.1 Approximately half of patients with early-stage disease have accelerated disease progression, whereas many of the remaining patients can survive for more than a decade without requiring therapy.2 The immunoglobulin heavy chain (IgVH) gene mutation status, leukemic cell expression of CD38 and 70-kD ζ-associated protein, and chromosome analysis by fluorescence in situ hybridization have all been useful in identifying patients with early-stage disease with biologically aggressive disease and shorter survival times.3-7 Unfortunately, all these new prognostic tests are expensive and require considerable technical exMayo Clin Proc.
•
pertise and equipment and thus are not available to many patients with CLL, especially those with limited resources or living in developing countries. Therefore, alternative, less expensive predictive tests are needed. Smudge cells are CLL cells ruptured during smear preparation. The interpatient variability in the percentage of smudge cells on a peripheral blood smear has long been recognized and is independent of the absolute lymphocyte count (ALC) and method of staining.8 Using a proteomic approach, we recently discovered that expression of the cytoskeletal protein vimentin is higher in leukemic cells from patients with CLL and unmutated CLL than in those with mutated IgVH genes.9 High vimentin expression is associated with progressive disease and a shortened time from diagnosis to initial treatment (TTT). Because vimentin increases lymphocyte rigidity,10 we hypothesized that leukemic cells with an increased vimentin content would be less likely to rupture. Thus, patients with CLL with increased vimentin expression would have a lower percentage of smudge cells on their peripheral blood smears compared with patients with low vimentin expression. Because decreased vimentin expression correlates with both mutated IgVH gene mutation status and favorable prognosis, we hypothesized that a higher percentage of smudge cells on routine peripheral blood smears would predict a better prognosis for patients with early-stage CLL. PATIENTS AND METHODS This study was approved by the Mayo Foundation Institutional Review Board in accordance with federal regulations From the Division of Hematology (G.S.N., T.D.S., T.G.C., C.S.Z., N.E.K.), Division of Hematopathology (J.D.H.), Cancer Center Statistics (S.M.G., B.R.L.), and Department of Immunology (D.F.J.), College of Medicine, Mayo Clinic, Rochester, Minn. This work was supported in part by grants from the National Institutes of Health and National Cancer Institute (CA92541) and philanthropic support from Mr Edson Spencer and the Donner family. Individual reprints of this article are not available. Address correspondence to Grzegorz S. Nowakowski, MD, Division of Hematology, College of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (e-mail: nowakowski [email protected]). © 2007 Mayo Foundation for Medical Education and Research
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
449
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
FIGURE 1. Peripheral blood smears of 4 patients with early-stage chronic lymphocytic leukemia. Upper left, patient with a low absolute lymphocyte count (ALC) (5.2 × 109/L) and a high percentage of smudge cells (37%). Arrowhead shows a typical smudge cell, and arrow shows a basket cell; upper right, patient with a low ALC and a low percentage of smudge cells (14%); lower left, patient with a high ALC (143 × 109/L) and a high percentage of smudge cells (44%); and lower right, patient with a high ALC (123 × 109/L) and a low percentage of smudge cells (11%) (Wright-Giemsa stain, original magnification ×400; Olympus AX70 Microscope, Olympus America Inc, Center Valley, Pa; Nikon DXM 1200 camera and Nikon ACT-1, v.2.62 software, Nikon Instruments Inc, Melville, NY).
and the Declaration of Helsinki. Patients seen at the Mayo Clinic in Rochester, Minn, between September 1989 and December 2000 who were diagnosed as having early- and intermediate-stage (Rai stage 0-II) CLL, consented to be included in a prospective clinical database, and had peripheral blood smears archived before the initiation of treatment were included in this study. Criteria for diagnosis and indication of treatment adhered to the National Cancer Institute Working Group guidelines.11 PERIPHERAL BLOOD SMEAR EXAMINATION Archived Wright-Giemsa–stained blood smears originally obtained for clinical purposes were reviewed. The smear obtained closest to the date of diagnosis was used for each patient. In 15 patients, a second archived specimen was also evaluated to determine the variability of the percentage of smudge cells over time. The blood smears were prepared from either EDTA anticoagulated blood or a fin450
Mayo Clin Proc.
•
ger stick. All blood smears were prepared using a semiautomatic device used by our laboratory since the 1980s (Miniprep; Sedona Lab Products, Sedona, Ariz) in which a simple spring mechanism pulls a drop of blood along a slide. According to established criteria, smudge cells were defined as broken cells with no intact cytoplasm and a disrupted nuclear membrane (Figure 1).12 A total of 200 lymphocytes and smudge cells were counted on each slide and the results expressed as a percentage of the total lymphocytes (intact and smudged). Because some other laboratories use the manual wedge method, in which a drop of blood is manually pulled by a glass cover, or the Autoprep (Sedona Lab Products) semiautomatic method (essentially the same procedure as the Miniprep), we also freshly prepared peripheral blood smears from 10 patients with CLL using each of these methods to allow comparison of smudging using these different techniques. In addition, to determine the reproducibility and
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
consistency of smear smudge cell counts among different observers, peripheral smears were made and read in a blinded manner by 3 hematopathology technologists. VIMENTIN EXPRESSION BY CLL CELLS Vimentin expression in peripheral blood CLL cells was measured by flow cytometry. Briefly, mononuclear cells isolated by density gradient centrifugation were stained with monoclonal phycoerythrin–labeled anti-CD19 antibodies (BD Pharmingen, San Jose, Calif), permeabilized (Fix and Perm Solution; Caltag Laboratories, Burlingame, Calif), and stained with antivimentin fluorescein isothiocyanate–labeled monoclonal antibody RV202 (Santa Cruz Biotechnology, Santa Cruz, Calif). The vimentin content was assessed in CD19+ gated cells by flow cytometry (FACScan; BD Biosciences, San Jose, Calif). Mean fluorescence intensity (MFI) of vimentin was compensated for the isotype control MFI. ASSESSMENT OF IGVH GENE MUTATIONAL STATUS The IgVH gene mutation status was assessed as previously described.13 Patients were classified as having unmutated status if there was 98% to 100% homology with the germline IgVH gene sequence and mutated status if there was less than 98% IgVH gene homology.
TABLE 1. Clinical Characteristics of the 75 Study Patients* Characteristic
No. (%)
Male Female Median age (y) (range) Rai stage 0 I II IgVH gene mutation status† Mutated Unmutated Median ALC at the time of smear (× 109/L) (range)
46 (61) 29 (39) 63.6 (33-86)
Sex
56 (75) 13 (17) 6 (8) 29 (64) 16 (36) 16.8 (5.3-165.7)
*ALC = absolute lymphocyte count; IgVH = immunoglobulin heavy chain. †Available for 45 patients.
median TTT for patients with unmutated IgVH status was 64.2 months (95% confidence interval [CI], 40.6-100.3 months) and was not reached for patients with mutated IgVH status (95% CI, 95.3 months to not yet reached; P=.002). The median survival for the entire cohort was not reached; the predicted 10-year overall survival rate was 77.2% (95% CI, 64%-93%).
STATISTICAL ANALYSES The differences between percentages of smudge cells on peripheral blood smears generated by using different methods and by independent observers were compared using the tables for calculating the least significant differences for leukocyte differential counts with 200 cells and a 2-sided level of significance of 0.05 as developed by Rumke et al.14 The nonparametric Spearman correlation test was used to assess correlations among continuous variables. The TTT was estimated using the Kaplan-Meier method, and differences between groups were assessed using the log-rank test.15 Univariate and multivariate analyses were performed using the Cox proportional hazard regression model.16 To discriminate between those who had a short TTT and those who had a longer TTT, a categorical variable for the percentage of smudge cells was defined using a minimum P value approach to identify the optimal cutpoint.
PERIPHERAL BLOOD SMEAR The median time from the date of diagnosis to the first available smear was 22 months (range, 0-162 months), and 39 patients (52%) had smears available for review within 24 months of diagnosis. In 15 patients, blood smear slides obtained at diagnosis were compared with slides made at later time points after diagnosis (median time between slides, 48 months; range, 9-72 months). The percentage of smudge cells had not changed significantly over time (P=.61). The median ALC at the time the peripheral blood smear was made was 16.8 × 109/L (range, 5.3-165.7 × 109/L). The median percentage of smudge cells was 27% (range, 4%72%). No correlation was found between the percentage of smudge cells and the ALC (r=0.04; P=.73). Concordance was found between slides prepared using the Miniprep device and the Autoprep device in 29 (97%) of 30 readouts (Table 2).14 In contrast, slides prepared by the manual wedge method were concordant in 17 (57%) of 30 readouts. A 93% (28/30 readouts) concordance rate was
RESULTS
TABLE 2. Concordance of Smudge Cell Percentage Prepared Using Miniprep With Autoprep and the Manual Wedge Method*
The clinical characteristics of the 75 eligible patients with blood smears are given in Table 1. The median follow-up from diagnosis was 75 months (range, 9-202 months). Thirty-two patients (43%) had been treated as of last follow-up with a median TTT of 95.3 months. The IgVH gene mutational status was available for 45 patients (60%). The Mayo Clin Proc.
•
Reader
Autoprep
Wedge method
1 2 3
10/10 9/10 10/10
5/10 6/10 6/10
*Data are presented as concordant/total readouts. The least significant differences (α=.05) were estimated for a 200-cell differential as described by Rumke et al.14
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
451
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
TABLE 3. Multivariate Analysis of the Impact of IgVH Gene Mutational Status and Percentage of Smudge Cells on Time From Diagnosis to Initial Treatment in Early-Stage Chronic Lymphocytic Leukemia Adjusted for Rai Stage* HR (95% CI)
P value
Rai stage of I or II at diagnosis Smudge cells ≥30% Mutated IgVH gene status
1.98 (0.82-4.79) 0.17 (0.05-0.56) 0.21 (0.07-0.60)
.13 .003 .004
Untreated patients (%)
Risk group
100
*CI = confidence interval; HR = hazard ratio; IgVH = immunoglobulin heavy chain.
80 60 40 20 0
found in the estimated percentage of smudge cells across the readers for the Miniprep and Autoprep methods, whereas concordance between readers when the manual wedge method was used was 57% (17/30 readouts).
0
DISCUSSION In 1896, Gumprecht17 described the presence of smudge cells (also known as shadows of Gumprecht or basket cells) 452
Mayo Clin Proc.
•
4
6
8
10
12
8
10
12
8
10
12
Time (y)
Untreated patients (%)
100
RELATIONSHIP AMONG THE PERCENTAGE OF SMUDGE CELLS, VIMENTIN EXPRESSION, AND IGVH GENE MUTATION STATUS We studied a subset of 21 patients with CLL who had pretreatment cryopreserved samples available for analysis. Cells from all 21 patients expressed vimentin with an MFI of 383 (range, 211-1666). Vimentin expression was inversely correlated with the percentage of smudge cells (r=–0.57; P=.007). The percentage of smudge cells was significantly higher in patients with mutated IgVH status (median, 31%) than in patients with unmutated IgVH status (median, 13%) (P=.02).
80 60 40 ≥30% Smudge cells <30% Smudge cells
20 0 0
2
4
6
Time (y) 100
Untreated patients (%)
PERCENTAGE OF SMUDGE CELLS AND CLINICAL OUTCOME The percentage of smudge cells as a continuous variable was associated with both a prolonged TTT (P=.002) and longer overall survival (P=.04). Using a minimum P value approach, we established an optimal cutpoint of 30% to define a categorical variable for the percentage of smudge cells. Patients with less than 30% smudge cells had a significantly shorter median TTT (72.7 months; 95% CI, 48.5-92.2 months) compared with patients with 30% or more smudge cells (median TTT, not yet reached; 95% CI, 100.3 months to not yet reached; P=.001) (Figure 2, middle). The difference between the groups remained statistically significant for a subset of patients with Rai stage 0 disease only (Figure 2, bottom). The estimated percentage of all patients with less than 30% smudge cells who required treatment within 10 years of diagnosis was 77% compared with 37% of patients with 30% or more smudge cells (P=.001). In a multivariate Cox model adjusted for Rai stage, the percentage of smudge cells (<30% vs ≥30%) and IgVH gene mutation status remained significant prognostic factors (Table 3).
2
80 60 40 ≥30% Smudge cells <30% Smudge cells
20 0 0
2
4
6
Time (y)
FIGURE 2. Kaplan-Meier plots show time from diagnosis to initial treatment (TTT) for all 75 patients (top) and for patients with 30% or more smudge cells vs those with less than 30% smudge cells for all patients (middle) (log rank P=.001) and patients with Rai stage 0 disease only (bottom) (log rank P=.001). The median TTT for the entire cohort was 95.3 months. Patients with less than 30% smudge cells on routine blood smears had a median TTT of 72.7 months, whereas the median TTT for patients with 30% or more smudge cells was not reached (P=.001). For patients with Rai stage 0 disease, median TTT was 76.3 months for patients with less than 30% smudge cells, and the median TTT was not reached for patients with 30% or more smudge cells (P=.01).
on the blood smears of patients with “lymphocytic leukemia.” The formation of smudge cells was initially considered an unimportant artifact of slide preparation, reflecting the “fragility” of CLL cells.18 In 1959 Heinivaara8 made 2 intriguing observations: first, that the percentage of smudge cells was not dependent simply on the degree of lymphocy-
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
tosis or the slide stain method, and second, that “smudging” appeared to be patient specific. In 1993, Binet et al19 described a correlation between the presence of dense chromatin and the number of smudge cells and postulated that smudge cells are from lymphocytes on the “verge of apoptosis.” We recently reported that a high expression of the cytoskeleton protein vimentin is associated with the presence of an unmutated IgVH gene and shortened TTT.9 Because vimentin is critical for lymphocyte rigidity, low vimentin expression could render cells more “fragile” and susceptible to smudging when making peripheral blood smears. We have now shown that low vimentin expression correlates with a high percentage of smudge cells on the peripheral blood smear. In addition to its mechanical role, vimentin is involved in signal transduction and cell activation of normal and malignant cells.20-22 Vimentin expression is associated with adverse prognosis in patients with solid tumors.22-24 In the current study, we demonstrate that the percentage of smudge cells on a peripheral blood smear is lower in patients with unmutated IgVH genes and that the percentage of smudge cells as a continuous variable correlates with both TTT and overall survival. The cutoff of 30% smudge cells was identified on the basis of the minimal P value approach and stratified patients into 2 groups with significantly different TTT. Importantly, similar to other investigators, we found no correlation between ALC and the percentage of smudge cells, and the percentage of smudge cells in patients did not change significantly over time. These findings indicate that the percentage of smudge cells is not simply a surrogate marker of disease burden but reflects differences in the biologic characteristics of the leukemic cells. Simple semiautomatic devices (Miniprep and Autoprep) and the manual wedge method are routinely used to prepare slides in hematopathology laboratories. In our study, generation of blood smears with semiautomatic devices yielded consistent smudge cell percentages; however, significant differences in intraobserver and interobserver counts were seen when the manual wedge method was used. This finding is consistent with previous findings, likely reflecting the nonuniform spread with the manual method. The reproducibility of a white cell differential count can be improved by counting more cells per slide.25 Whether counting more than 200 cells to estimate the percentage of smudge cells would improve reproducibility of all the methods used remains to be established. Currently, we recommend using slides prepared by semiautomatic methods and a minimum of a 200-cell count for future studies. CONCLUSION Our finding that the high percentage of smudge cells on a peripheral blood smear is associated with mutated IgVH Mayo Clin Proc.
•
gene status, a longer TTT, and better overall survival in patients with early-stage CLL could provide a simple and inexpensive prognostic test available to nearly all patients worldwide with a diagnosis of CLL. REFERENCES 1. Call TG, Phyliky RL, Noel P, et al. Incidence of chronic lymphocytic leukemia in Olmsted County, Minnesota, 1935 through 1989, with emphasis on changes in initial stage at diagnosis. Mayo Clin Proc. 1994;69:323-328. 2. Chiorazzi N, Rai KR, Ferrarini M. Chronic lymphocytic leukemia. N Engl J Med. 2005;352:804-815. 3. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood. 1999;94:1848-1854. 4. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94:1840-1847. 5. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. N Engl J Med. 2003;348:1764-1775. 6. Rassenti LZ, Huynh L, Toy TL, et al. ZAP-70 compared with immunoglobulin heavy-chain gene mutation status as a predictor of disease progression in chronic lymphocytic leukemia. N Engl J Med. 2004;351:893-901. 7. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343:1910-1916. 8. Heinivaara O. Smudge cells in lymphocytic leukemia. Ann Med Intern Fenn. 1959;48:69-75. 9. Nowakowski GS, Lee YK, Bone ND, et al. Analysis of chronic lymphocytic leukemia cells identifies vimentin as a novel prognostic factor for aggressive disease. Blood. 2005;106(pt 1):209a. Abstract 707. 10. Brown MJ, Hallam JA, Colucci-Guyon E, Shaw S. Rigidity of circulating lymphocytes is primarily conferred by vimentin intermediate filaments. J Immunol. 2001;166:6640-6646. 11. Cheson BD, Bennett JM, Grever M, et al. National Cancer Institutesponsored Working Group guidelines for chronic lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood. 1996;87:4990-4997. 12. Henry BJ, ed. Clinical Diagnosis and Management by Laboratory Methods, 20th ed. Philadelphia, Pa: WB Saunders; 2001. 13. Jelinek DF, Tschumper RC, Geyer SM, et al. Analysis of clonal B-cell CD38 and immunoglobulin variable region sequence status in relation to clinical outcome for B-chronic lymphocytic leukaemia. Br J Haematol. 2001;115:854-861. 14. Rumke CL, Bezemer PD, Kuik DJ. Normal values and least significant differences for differential leukocyte counts. J Chronic Dis. 1975;28:661-668. 15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. 16. Cox DR. Regression models and life-tables. J R Stat Soc [B]. 1972;34:187220. 17. Gumprecht F. Leukozytenzerfall im Blute bei Leukamie und bei schweren Anamien. Dtsch Arch Klin Med. 1896;57:523-548. 18. Macdonald D, Richardson H, Raby A. Practice guidelines on the reporting of smudge cells in the white blood cell differential count. Arch Pathol Lab Med. 2003;127:105. 19. Binet JL, Baudet S, Mentz F, et al. Basket cells or shadow cells of Gumprecht: a scanning electron microscope study, and the correlation between percentages of basket cells, and cells with altered chromatin structure (dense cells), in chronic lymphocytic leukemia. Blood Cells. 1993;19:573-581. 20. Nieminen M, Henttinen T, Merinen M, Marttila-Ichihara F, Eriksson JE, Jalkanen S. Vimentin function in lymphocyte adhesion and transcellular migration. Nat Cell Biol. 2006 Feb;8:156-162. Epub 2006 Jan 22. 21. Wang R, Li QF, Anfinogenova Y, Tang DD. Dissociation of Crkassociated substrate from the vimentin network is regulated by p21-activated kinase on ACh activation of airway smooth muscle. Am J Physiol Lung Cell Mol Physiol. 2007 Jan;292:L240-L248. Epub 2006 Sep 22. 22. Singh S, Sadacharan S, Su S, Belldegrun A, Persad S, Singh G. Overexpression of vimentin: role in the invasive phenotype in an androgenindependent model of prostate cancer. Cancer Res. 2003;63:2306-2311. 23. Sommers CL, Heckford SE, Skerker JM, et al. Loss of epithelial markers and acquisition of vimentin expression in adriamycin- and vinblastine-resistant human breast cancer cell lines. Cancer Res. 1992;52:5190-5197. 24. Thomas PA, Kirschmann DA, Cerhan JR, et al. Association between keratin and vimentin expression, malignant phenotype, and survival in postmenopausal breast cancer patients. Clin Cancer Res. 1999;5:2698-2703. 25. Rumke CL. Significance of the differential leukocyte count for the demonstration and the exclusion of the presence of atypical cells [in Dutch]. Ned Tijdschr Geneeskd. 1985;129:940-943.
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
453
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
Using Smudge Cells on Routine Blood Smears to Predict Clinical Outcome in Chronic Lymphocytic Leukemia: A Universally Available Prognostic Test GRZEGORZ S. NOWAKOWSKI, MD; JAMES D. HOYER, MD; TAIT D. SHANAFELT, MD; SUSAN M. GEYER, PHD; BETSY R. LAPLANT, MS; TIMOTHY G. CALL, MD; DIANE F. JELINEK, PHD; CLIVE S. ZENT, MD; AND NEIL E. KAY, MD Recently developed prognostic tests in early Rai and Binet stage chronic lymphocytic leukemia (CLL) require considerable technologic expertise and are not available worldwide. Smudge cells are CLL cells ruptured during smear preparation. We hypothesized that smudge cell formation is inversely correlated with expression of vimentin, a cytoskeletal protein and prognostic marker, and that the percentage of smudge cells would predict prognosis in CLL. We reviewed the blood smears of 75 patients with previously untreated early- and intermediate-stage CLL (Rai stage 0-II) who were seen at the Mayo Clinic in Rochester, Minn, between September 1989 and December 2000. A total of 200 lymphocytes and smudge cells were counted on each slide and the results expressed as a percentage of the total lymphocytes (intact and smudged). The median percentage of smudge cells was 27% (range, 4%-72%). The percentage of smudge cells inversely correlated with vimentin expression (r=–0.57; P=.007). The median percentage of smudge cells was higher in patients with the mutated immunoglobulin heavy chain gene than in those with the unmutated immunoglobulin heavy chain gene (31% vs 13%; P=.02). Patients with less than 30% smudge cells had a median time from diagnosis to initial treatment of 72.7 months, whereas the median time from diagnosis to initial treatment in patients with 30% or more smudge cells was not reached (P=.001). The percentage of smudge cells as a continuous variable correlated with overall survival (P=.04). The estimation of smudge cells on a blood smear could be a universally available prognostic test in early-stage CLL.
Mayo Clin Proc. 2007;82(4):449-453 ALC = absolute lymphocyte count; CI = confidence interval; CLL = chronic lymphocytic leukemia; IgVH = immunoglobulin heavy chain; MFI = mean fluorescence intensity; TTT = time from diagnosis to initial treatment
W
ith the widespread use of automated blood cell counters and flow cytometric immunophenotyping, 60% to 80% of patients with chronic lymphocytic leukemia (CLL) now have their condition diagnosed at an early stage of disease.1 Approximately half of patients with early-stage disease have accelerated disease progression, whereas many of the remaining patients can survive for more than a decade without requiring therapy.2 The immunoglobulin heavy chain (IgVH) gene mutation status, leukemic cell expression of CD38 and 70-kD ζ-associated protein, and chromosome analysis by fluorescence in situ hybridization have all been useful in identifying patients with early-stage disease with biologically aggressive disease and shorter survival times.3-7 Unfortunately, all these new prognostic tests are expensive and require considerable technical exMayo Clin Proc.
•
pertise and equipment and thus are not available to many patients with CLL, especially those with limited resources or living in developing countries. Therefore, alternative, less expensive predictive tests are needed. Smudge cells are CLL cells ruptured during smear preparation. The interpatient variability in the percentage of smudge cells on a peripheral blood smear has long been recognized and is independent of the absolute lymphocyte count (ALC) and method of staining.8 Using a proteomic approach, we recently discovered that expression of the cytoskeletal protein vimentin is higher in leukemic cells from patients with CLL and unmutated CLL than in those with mutated IgVH genes.9 High vimentin expression is associated with progressive disease and a shortened time from diagnosis to initial treatment (TTT). Because vimentin increases lymphocyte rigidity,10 we hypothesized that leukemic cells with an increased vimentin content would be less likely to rupture. Thus, patients with CLL with increased vimentin expression would have a lower percentage of smudge cells on their peripheral blood smears compared with patients with low vimentin expression. Because decreased vimentin expression correlates with both mutated IgVH gene mutation status and favorable prognosis, we hypothesized that a higher percentage of smudge cells on routine peripheral blood smears would predict a better prognosis for patients with early-stage CLL. PATIENTS AND METHODS This study was approved by the Mayo Foundation Institutional Review Board in accordance with federal regulations From the Division of Hematology (G.S.N., T.D.S., T.G.C., C.S.Z., N.E.K.), Division of Hematopathology (J.D.H.), Cancer Center Statistics (S.M.G., B.R.L.), and Department of Immunology (D.F.J.), College of Medicine, Mayo Clinic, Rochester, Minn. This work was supported in part by grants from the National Institutes of Health and National Cancer Institute (CA92541) and philanthropic support from Mr Edson Spencer and the Donner family. Individual reprints of this article are not available. Address correspondence to Grzegorz S. Nowakowski, MD, Division of Hematology, College of Medicine, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (e-mail: nowakowski [email protected]). © 2007 Mayo Foundation for Medical Education and Research
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
449
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
FIGURE 1. Peripheral blood smears of 4 patients with early-stage chronic lymphocytic leukemia. Upper left, patient with a low absolute lymphocyte count (ALC) (5.2 × 109/L) and a high percentage of smudge cells (37%). Arrowhead shows a typical smudge cell, and arrow shows a basket cell; upper right, patient with a low ALC and a low percentage of smudge cells (14%); lower left, patient with a high ALC (143 × 109/L) and a high percentage of smudge cells (44%); and lower right, patient with a high ALC (123 × 109/L) and a low percentage of smudge cells (11%) (Wright-Giemsa stain, original magnification ×400; Olympus AX70 Microscope, Olympus America Inc, Center Valley, Pa; Nikon DXM 1200 camera and Nikon ACT-1, v.2.62 software, Nikon Instruments Inc, Melville, NY).
and the Declaration of Helsinki. Patients seen at the Mayo Clinic in Rochester, Minn, between September 1989 and December 2000 who were diagnosed as having early- and intermediate-stage (Rai stage 0-II) CLL, consented to be included in a prospective clinical database, and had peripheral blood smears archived before the initiation of treatment were included in this study. Criteria for diagnosis and indication of treatment adhered to the National Cancer Institute Working Group guidelines.11 PERIPHERAL BLOOD SMEAR EXAMINATION Archived Wright-Giemsa–stained blood smears originally obtained for clinical purposes were reviewed. The smear obtained closest to the date of diagnosis was used for each patient. In 15 patients, a second archived specimen was also evaluated to determine the variability of the percentage of smudge cells over time. The blood smears were prepared from either EDTA anticoagulated blood or a fin450
Mayo Clin Proc.
•
ger stick. All blood smears were prepared using a semiautomatic device used by our laboratory since the 1980s (Miniprep; Sedona Lab Products, Sedona, Ariz) in which a simple spring mechanism pulls a drop of blood along a slide. According to established criteria, smudge cells were defined as broken cells with no intact cytoplasm and a disrupted nuclear membrane (Figure 1).12 A total of 200 lymphocytes and smudge cells were counted on each slide and the results expressed as a percentage of the total lymphocytes (intact and smudged). Because some other laboratories use the manual wedge method, in which a drop of blood is manually pulled by a glass cover, or the Autoprep (Sedona Lab Products) semiautomatic method (essentially the same procedure as the Miniprep), we also freshly prepared peripheral blood smears from 10 patients with CLL using each of these methods to allow comparison of smudging using these different techniques. In addition, to determine the reproducibility and
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
consistency of smear smudge cell counts among different observers, peripheral smears were made and read in a blinded manner by 3 hematopathology technologists. VIMENTIN EXPRESSION BY CLL CELLS Vimentin expression in peripheral blood CLL cells was measured by flow cytometry. Briefly, mononuclear cells isolated by density gradient centrifugation were stained with monoclonal phycoerythrin–labeled anti-CD19 antibodies (BD Pharmingen, San Jose, Calif), permeabilized (Fix and Perm Solution; Caltag Laboratories, Burlingame, Calif), and stained with antivimentin fluorescein isothiocyanate–labeled monoclonal antibody RV202 (Santa Cruz Biotechnology, Santa Cruz, Calif). The vimentin content was assessed in CD19+ gated cells by flow cytometry (FACScan; BD Biosciences, San Jose, Calif). Mean fluorescence intensity (MFI) of vimentin was compensated for the isotype control MFI. ASSESSMENT OF IGVH GENE MUTATIONAL STATUS The IgVH gene mutation status was assessed as previously described.13 Patients were classified as having unmutated status if there was 98% to 100% homology with the germline IgVH gene sequence and mutated status if there was less than 98% IgVH gene homology.
TABLE 1. Clinical Characteristics of the 75 Study Patients* Characteristic
No. (%)
Male Female Median age (y) (range) Rai stage 0 I II IgVH gene mutation status† Mutated Unmutated Median ALC at the time of smear (× 109/L) (range)
46 (61) 29 (39) 63.6 (33-86)
Sex
56 (75) 13 (17) 6 (8) 29 (64) 16 (36) 16.8 (5.3-165.7)
*ALC = absolute lymphocyte count; IgVH = immunoglobulin heavy chain. †Available for 45 patients.
median TTT for patients with unmutated IgVH status was 64.2 months (95% confidence interval [CI], 40.6-100.3 months) and was not reached for patients with mutated IgVH status (95% CI, 95.3 months to not yet reached; P=.002). The median survival for the entire cohort was not reached; the predicted 10-year overall survival rate was 77.2% (95% CI, 64%-93%).
STATISTICAL ANALYSES The differences between percentages of smudge cells on peripheral blood smears generated by using different methods and by independent observers were compared using the tables for calculating the least significant differences for leukocyte differential counts with 200 cells and a 2-sided level of significance of 0.05 as developed by Rumke et al.14 The nonparametric Spearman correlation test was used to assess correlations among continuous variables. The TTT was estimated using the Kaplan-Meier method, and differences between groups were assessed using the log-rank test.15 Univariate and multivariate analyses were performed using the Cox proportional hazard regression model.16 To discriminate between those who had a short TTT and those who had a longer TTT, a categorical variable for the percentage of smudge cells was defined using a minimum P value approach to identify the optimal cutpoint.
PERIPHERAL BLOOD SMEAR The median time from the date of diagnosis to the first available smear was 22 months (range, 0-162 months), and 39 patients (52%) had smears available for review within 24 months of diagnosis. In 15 patients, blood smear slides obtained at diagnosis were compared with slides made at later time points after diagnosis (median time between slides, 48 months; range, 9-72 months). The percentage of smudge cells had not changed significantly over time (P=.61). The median ALC at the time the peripheral blood smear was made was 16.8 × 109/L (range, 5.3-165.7 × 109/L). The median percentage of smudge cells was 27% (range, 4%72%). No correlation was found between the percentage of smudge cells and the ALC (r=0.04; P=.73). Concordance was found between slides prepared using the Miniprep device and the Autoprep device in 29 (97%) of 30 readouts (Table 2).14 In contrast, slides prepared by the manual wedge method were concordant in 17 (57%) of 30 readouts. A 93% (28/30 readouts) concordance rate was
RESULTS
TABLE 2. Concordance of Smudge Cell Percentage Prepared Using Miniprep With Autoprep and the Manual Wedge Method*
The clinical characteristics of the 75 eligible patients with blood smears are given in Table 1. The median follow-up from diagnosis was 75 months (range, 9-202 months). Thirty-two patients (43%) had been treated as of last follow-up with a median TTT of 95.3 months. The IgVH gene mutational status was available for 45 patients (60%). The Mayo Clin Proc.
•
Reader
Autoprep
Wedge method
1 2 3
10/10 9/10 10/10
5/10 6/10 6/10
*Data are presented as concordant/total readouts. The least significant differences (α=.05) were estimated for a 200-cell differential as described by Rumke et al.14
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
451
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
TABLE 3. Multivariate Analysis of the Impact of IgVH Gene Mutational Status and Percentage of Smudge Cells on Time From Diagnosis to Initial Treatment in Early-Stage Chronic Lymphocytic Leukemia Adjusted for Rai Stage* HR (95% CI)
P value
Rai stage of I or II at diagnosis Smudge cells ≥30% Mutated IgVH gene status
1.98 (0.82-4.79) 0.17 (0.05-0.56) 0.21 (0.07-0.60)
.13 .003 .004
Untreated patients (%)
Risk group
100
*CI = confidence interval; HR = hazard ratio; IgVH = immunoglobulin heavy chain.
80 60 40 20 0
found in the estimated percentage of smudge cells across the readers for the Miniprep and Autoprep methods, whereas concordance between readers when the manual wedge method was used was 57% (17/30 readouts).
0
DISCUSSION In 1896, Gumprecht17 described the presence of smudge cells (also known as shadows of Gumprecht or basket cells) 452
Mayo Clin Proc.
•
4
6
8
10
12
8
10
12
8
10
12
Time (y)
Untreated patients (%)
100
RELATIONSHIP AMONG THE PERCENTAGE OF SMUDGE CELLS, VIMENTIN EXPRESSION, AND IGVH GENE MUTATION STATUS We studied a subset of 21 patients with CLL who had pretreatment cryopreserved samples available for analysis. Cells from all 21 patients expressed vimentin with an MFI of 383 (range, 211-1666). Vimentin expression was inversely correlated with the percentage of smudge cells (r=–0.57; P=.007). The percentage of smudge cells was significantly higher in patients with mutated IgVH status (median, 31%) than in patients with unmutated IgVH status (median, 13%) (P=.02).
80 60 40 ≥30% Smudge cells <30% Smudge cells
20 0 0
2
4
6
Time (y) 100
Untreated patients (%)
PERCENTAGE OF SMUDGE CELLS AND CLINICAL OUTCOME The percentage of smudge cells as a continuous variable was associated with both a prolonged TTT (P=.002) and longer overall survival (P=.04). Using a minimum P value approach, we established an optimal cutpoint of 30% to define a categorical variable for the percentage of smudge cells. Patients with less than 30% smudge cells had a significantly shorter median TTT (72.7 months; 95% CI, 48.5-92.2 months) compared with patients with 30% or more smudge cells (median TTT, not yet reached; 95% CI, 100.3 months to not yet reached; P=.001) (Figure 2, middle). The difference between the groups remained statistically significant for a subset of patients with Rai stage 0 disease only (Figure 2, bottom). The estimated percentage of all patients with less than 30% smudge cells who required treatment within 10 years of diagnosis was 77% compared with 37% of patients with 30% or more smudge cells (P=.001). In a multivariate Cox model adjusted for Rai stage, the percentage of smudge cells (<30% vs ≥30%) and IgVH gene mutation status remained significant prognostic factors (Table 3).
2
80 60 40 ≥30% Smudge cells <30% Smudge cells
20 0 0
2
4
6
Time (y)
FIGURE 2. Kaplan-Meier plots show time from diagnosis to initial treatment (TTT) for all 75 patients (top) and for patients with 30% or more smudge cells vs those with less than 30% smudge cells for all patients (middle) (log rank P=.001) and patients with Rai stage 0 disease only (bottom) (log rank P=.001). The median TTT for the entire cohort was 95.3 months. Patients with less than 30% smudge cells on routine blood smears had a median TTT of 72.7 months, whereas the median TTT for patients with 30% or more smudge cells was not reached (P=.001). For patients with Rai stage 0 disease, median TTT was 76.3 months for patients with less than 30% smudge cells, and the median TTT was not reached for patients with 30% or more smudge cells (P=.01).
on the blood smears of patients with “lymphocytic leukemia.” The formation of smudge cells was initially considered an unimportant artifact of slide preparation, reflecting the “fragility” of CLL cells.18 In 1959 Heinivaara8 made 2 intriguing observations: first, that the percentage of smudge cells was not dependent simply on the degree of lymphocy-
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
PROGNOSTIC SIGNIFICANCE OF SMUDGE CELLS IN CLL
tosis or the slide stain method, and second, that “smudging” appeared to be patient specific. In 1993, Binet et al19 described a correlation between the presence of dense chromatin and the number of smudge cells and postulated that smudge cells are from lymphocytes on the “verge of apoptosis.” We recently reported that a high expression of the cytoskeleton protein vimentin is associated with the presence of an unmutated IgVH gene and shortened TTT.9 Because vimentin is critical for lymphocyte rigidity, low vimentin expression could render cells more “fragile” and susceptible to smudging when making peripheral blood smears. We have now shown that low vimentin expression correlates with a high percentage of smudge cells on the peripheral blood smear. In addition to its mechanical role, vimentin is involved in signal transduction and cell activation of normal and malignant cells.20-22 Vimentin expression is associated with adverse prognosis in patients with solid tumors.22-24 In the current study, we demonstrate that the percentage of smudge cells on a peripheral blood smear is lower in patients with unmutated IgVH genes and that the percentage of smudge cells as a continuous variable correlates with both TTT and overall survival. The cutoff of 30% smudge cells was identified on the basis of the minimal P value approach and stratified patients into 2 groups with significantly different TTT. Importantly, similar to other investigators, we found no correlation between ALC and the percentage of smudge cells, and the percentage of smudge cells in patients did not change significantly over time. These findings indicate that the percentage of smudge cells is not simply a surrogate marker of disease burden but reflects differences in the biologic characteristics of the leukemic cells. Simple semiautomatic devices (Miniprep and Autoprep) and the manual wedge method are routinely used to prepare slides in hematopathology laboratories. In our study, generation of blood smears with semiautomatic devices yielded consistent smudge cell percentages; however, significant differences in intraobserver and interobserver counts were seen when the manual wedge method was used. This finding is consistent with previous findings, likely reflecting the nonuniform spread with the manual method. The reproducibility of a white cell differential count can be improved by counting more cells per slide.25 Whether counting more than 200 cells to estimate the percentage of smudge cells would improve reproducibility of all the methods used remains to be established. Currently, we recommend using slides prepared by semiautomatic methods and a minimum of a 200-cell count for future studies. CONCLUSION Our finding that the high percentage of smudge cells on a peripheral blood smear is associated with mutated IgVH Mayo Clin Proc.
•
gene status, a longer TTT, and better overall survival in patients with early-stage CLL could provide a simple and inexpensive prognostic test available to nearly all patients worldwide with a diagnosis of CLL. REFERENCES 1. Call TG, Phyliky RL, Noel P, et al. Incidence of chronic lymphocytic leukemia in Olmsted County, Minnesota, 1935 through 1989, with emphasis on changes in initial stage at diagnosis. Mayo Clin Proc. 1994;69:323-328. 2. Chiorazzi N, Rai KR, Ferrarini M. Chronic lymphocytic leukemia. N Engl J Med. 2005;352:804-815. 3. Hamblin TJ, Davis Z, Gardiner A, Oscier DG, Stevenson FK. Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood. 1999;94:1848-1854. 4. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood. 1999;94:1840-1847. 5. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as a surrogate for immunoglobulin-variable-region mutations in chronic lymphocytic leukemia. N Engl J Med. 2003;348:1764-1775. 6. Rassenti LZ, Huynh L, Toy TL, et al. ZAP-70 compared with immunoglobulin heavy-chain gene mutation status as a predictor of disease progression in chronic lymphocytic leukemia. N Engl J Med. 2004;351:893-901. 7. Dohner H, Stilgenbauer S, Benner A, et al. Genomic aberrations and survival in chronic lymphocytic leukemia. N Engl J Med. 2000;343:1910-1916. 8. Heinivaara O. Smudge cells in lymphocytic leukemia. Ann Med Intern Fenn. 1959;48:69-75. 9. Nowakowski GS, Lee YK, Bone ND, et al. Analysis of chronic lymphocytic leukemia cells identifies vimentin as a novel prognostic factor for aggressive disease. Blood. 2005;106(pt 1):209a. Abstract 707. 10. Brown MJ, Hallam JA, Colucci-Guyon E, Shaw S. Rigidity of circulating lymphocytes is primarily conferred by vimentin intermediate filaments. J Immunol. 2001;166:6640-6646. 11. Cheson BD, Bennett JM, Grever M, et al. National Cancer Institutesponsored Working Group guidelines for chronic lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood. 1996;87:4990-4997. 12. Henry BJ, ed. Clinical Diagnosis and Management by Laboratory Methods, 20th ed. Philadelphia, Pa: WB Saunders; 2001. 13. Jelinek DF, Tschumper RC, Geyer SM, et al. Analysis of clonal B-cell CD38 and immunoglobulin variable region sequence status in relation to clinical outcome for B-chronic lymphocytic leukaemia. Br J Haematol. 2001;115:854-861. 14. Rumke CL, Bezemer PD, Kuik DJ. Normal values and least significant differences for differential leukocyte counts. J Chronic Dis. 1975;28:661-668. 15. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. 16. Cox DR. Regression models and life-tables. J R Stat Soc [B]. 1972;34:187220. 17. Gumprecht F. Leukozytenzerfall im Blute bei Leukamie und bei schweren Anamien. Dtsch Arch Klin Med. 1896;57:523-548. 18. Macdonald D, Richardson H, Raby A. Practice guidelines on the reporting of smudge cells in the white blood cell differential count. Arch Pathol Lab Med. 2003;127:105. 19. Binet JL, Baudet S, Mentz F, et al. Basket cells or shadow cells of Gumprecht: a scanning electron microscope study, and the correlation between percentages of basket cells, and cells with altered chromatin structure (dense cells), in chronic lymphocytic leukemia. Blood Cells. 1993;19:573-581. 20. Nieminen M, Henttinen T, Merinen M, Marttila-Ichihara F, Eriksson JE, Jalkanen S. Vimentin function in lymphocyte adhesion and transcellular migration. Nat Cell Biol. 2006 Feb;8:156-162. Epub 2006 Jan 22. 21. Wang R, Li QF, Anfinogenova Y, Tang DD. Dissociation of Crkassociated substrate from the vimentin network is regulated by p21-activated kinase on ACh activation of airway smooth muscle. Am J Physiol Lung Cell Mol Physiol. 2007 Jan;292:L240-L248. Epub 2006 Sep 22. 22. Singh S, Sadacharan S, Su S, Belldegrun A, Persad S, Singh G. Overexpression of vimentin: role in the invasive phenotype in an androgenindependent model of prostate cancer. Cancer Res. 2003;63:2306-2311. 23. Sommers CL, Heckford SE, Skerker JM, et al. Loss of epithelial markers and acquisition of vimentin expression in adriamycin- and vinblastine-resistant human breast cancer cell lines. Cancer Res. 1992;52:5190-5197. 24. Thomas PA, Kirschmann DA, Cerhan JR, et al. Association between keratin and vimentin expression, malignant phenotype, and survival in postmenopausal breast cancer patients. Clin Cancer Res. 1999;5:2698-2703. 25. Rumke CL. Significance of the differential leukocyte count for the demonstration and the exclusion of the presence of atypical cells [in Dutch]. Ned Tijdschr Geneeskd. 1985;129:940-943.
April 2007;82(4):449-453
•
www.mayoclinicproceedings.com
For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.
453