The long-term clinical implications of clonal chromosomal abnormalities in newly diagnosed chronic phase chronic myeloid leukemia patients treated with imatinib mesylate

Cancer Genetics 205 (2012) 563e571

The long-term clinical implications of clonal chromosomal abnormalities in newly diagnosed chronic phase chronic myeloid leukemia patients treated with imatinib mesylate Sung-Eun Lee a,b, Soo Young Choi a, Ju-Hee Bang a, Soo-Hyun Kim a, Eun-jung Jang a, Ji-Young Byeun a, Jin Eok Park a, Hye-Rim Jeon a, Yun Jeong Oh a, Myungshin Kim c, Dong-Wook Kim a,b,* a b

Cancer Research Institute, The Catholic University of Korea, Seoul, Korea; Department of Hematology, Seoul St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, Seoul, Korea; c Department of Laboratory Medicine, Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, Korea The aim of this study was to evaluate the long-term clinical significance of an additional chromosomal abnormality (ACA), variant Philadelphia chromosome (vPh) at diagnosis, and newly developed other chromosomal abnormalities (OCA) in patients with chronic myeloid leukemia (CML) on imatinib (IM) therapy. Sequential cytogenetic data from 281 consecutive new chronic phase CML patients were analyzed. With a median follow-up of 78.6 months, the 22 patients with vPh (P Z 0.034) or ACA (P Z 0.034) at diagnosis had more events of IM failure than did the patients with a standard Ph. The 5-year overall survival (OS), event-free survival (EFS), and failure-free survival (FFS) rates for patients with vPh at diagnosis were 77.8%, 75.0%, and 53.3%, respectively; for patients with ACA at diagnosis, 100%, 66.3%, and 52.1%, respectively; and for patients with a standard Ph, 96.0%, 91.3%, and 83.7%, respectively. During IM therapy, eight patients developed an OCA, which had no impact on outcomes as a time-dependent covariate in our Cox proportional hazards regression models. This study showed that vPh was associated with poor OS and FFS and that ACA had adverse effects on EFS and FFS. In addition, no OCA, except monosomy 7, had any prognostic impact, suggesting that the development of OCA may not require a change in treatment strategy. Keywords Chromosomal abnormality, chronic myeloid leukemia, imatinib mesylate ª 2012 Elsevier Inc. All rights reserved.

The development of clonal chromosomal abnormalities (CAs) in chronic myeloid leukemia (CML) was considered a step toward increased genetic instability or progression to accelerated phase (AP) and blast crisis (BC) in the era before imatinib (IM). Cytogenetic clonal evolution occurred in approximately 50e80% of patients with BC CML and was associated with poor prognosis (1,2). However, CAs other than the standard Philadelphia chromosome (Ph) are

Received June 11, 2012; received in revised form September 13, 2012; accepted September 20, 2012. * Corresponding author. E-mail address: [email protected] 2210-7762/$ - see front matter ª 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.cancergen.2012.09.003

described in less than 10% of cases of chronic phase (CP) CML at diagnosis (3,4). As the first BCR-ABL1 tyrosine kinase inhibitor (TKI), IM has been used widely for the treatment of CML, and many studies have been conducted to evaluate the prognostic significance of CA in the era of IM therapy (3,5e9). To date, conflicting clinical impacts of CA on IM response have been reported in different studies. Marin et al. analyzed 224 CP CML patients who received IM as first-line therapy and showed that the presence of additional CA (ACA) in Phpositive (Phþ) cells, either at diagnosis or emerging during therapy, was associated with poor outcomes (10). Studies focusing on the variant Philadelphia chromosome (vPh) have reported contradictory findings; some have suggested its association with worse prognosis (11), whereas others have

564 found no impact on prognosis (6,12). Recently, Fabarius et al. reported a negative prognostic impact of major-route ACAs at diagnosis compared with that of standard Ph, vPh, -Y, or minor-route ACAs, which emphasizes the importance of the type of chromosome aberration (9). In addition, although the presence of other CA (OCA) in Ph-negative (Ph-) cells has been noted in 3.6e8.1% of IM-treated patients in previous studies, the prognostic impact of OCA in Ph- cells is still not clear (10,13e15). Jabbour et al. demonstrated that of 246 patients with a cytogenetic response to IM therapy, 21 patients (9%) developed 23 OCAs, with a median follow-up of 37 months, and these patients exhibited inferior progression-free survival (PFS) and overall survival (OS) compared with those of patients without OCA (13). These results are distinct from those of a study by Marin et al., which showed that OCAs were not associated with OS or PFS (10). The discrepancies in these findings may have been caused by heterogeneous disease status, differences in treatment modalities, short duration of follow-up, or the predominance of the clones involved. Moreover, the exact clinical impact of CA cannot be assessed based on previous OCA studies because of a lack of cytogenetic data at diagnosis and a lack of serial follow-up data in the same populations. Moreover, because most previous studies did not uniformly report the time of CA development, a study with sequential long-term evaluation in a homogenous patient group is needed. Considering the importance of clarifying the comprehensive long-term prognostic impacts of CA, we conducted this study to investigate the effects of OCA during IM therapy as well as those of ACA and vPh at diagnosis on long-term responses to IM therapy and survival in newly diagnosed CP CML patients.

Materials and methods Patient selection and response analyses We analyzed 281 consecutive patients who were newly diagnosed with CP CML at Seoul St. Mary Hospital between January 2001 and May 2009 and who received 400 mg IM once daily with no prior treatment. Cytogenetic and molecular analyses were performed as previously suggested in the European Leukemia Net (ELN) recommendations (16). To confirm the existence of CA in complete cytogenetic response (CCyR), we performed an additional cytogenetic assay in patients who continued to exhibit optimal response after more than five years of IM therapy. All clinical data, including hematologic, cytogenetic and molecular response records, were registered and maintained in the Asia CML Registry (ACR) database system before use in this study. This study was approved by the institutional review board and conducted in accordance with the Declaration of Helsinki.

S.-E. Lee et al.

Quantification of BCR-ABL1 transcripts To evaluate the molecular response to treatment, we performed real-time quantitative reverse transcriptasee polymerase chain reaction (RT-Q-PCR) to measure the BCR-ABL1 transcript copy number, as described previously, at 3-month intervals and then 6-month intervals after achieving major molecular response (MMR). The results were expressed using the international scale (IS). MMR was defined as a BCR-ABL1 transcript level of 0.1% or lower on the IS. Complete molecular response (MR4.5) was defined as a reduction in the BCR-ABL1 transcript level to 0.0032% or lower on the IS.

Definitions Disease phase and response were defined according to recent criteria (16). Cytogenetic alterations and clonality were defined according to the International System for Human Cytogenetic Nomenclature (17). CAs at diagnosis were classified as additional or variant according to cytogenetic alterations in Phþ cells. OS was calculated from the day of IM therapy initiation to death or last contact. FFS was counted from the initial day of IM therapy to death, progression to AP or BC, treatment failure according to ELN recommendations (16), or last contact, whichever came first. In addition, EFS was counted from the day of IM start to death, progression to AP or BC, treatment failure according to ELN recommendations (16), treatment discontinuation for any reason, or last contact, whichever came first.

Statistical analysis The main end points of this study were OS, EFS, and FFS. Survival curves for OS, EFS, and FFS were plotted using the KaplaneMeier method and were compared using the logrank test. CA at diagnosis and variables with a P value less than 0.05, as determined by univariate analyses, were included in the multivariate analyses using the Cox proportional hazard regression model. The proportionality assumption was tested by adding a time-dependent covariate for each factor, and the effects of OCA in Ph- cells on the preceding outcomes were assessed using a time-dependent covariate in the final multivariate model. The probabilities of CCyR, MMR, and MR4.5 were compared by the cumulative incidence procedure, where CCyR, two consecutive MMRs, and two consecutive MRs4.5 were the events of interest, and IM discontinuation, death, and disease progression were the competitors. Statistical analyses were performed with the Statistical Package for the Social Sciences (version 13.0; SPSS, Chicago, IL).

Results

Cytogenetic assays

The characteristics of CA at diagnosis

Routine cytogenetic analyses were performed using standard G-banding in bone marrow (BM) aspirates, and all cytogenetic responses were estimated based on analyses of more than 20 metaphase cells.

The median number of BM aspirations for cytogenetic examination was 4 (range, 1e28). Of the 281 patients, 259 patients (92%) with a median age of 41 years (range, 18e77 y) had only a standard Ph translocation, 13 patients

Clonal CA in CML after imatinib

565

(5%) had ACA, and nine patients (3%) had a vPh at diagnosis. Age, sex, and Sokal score were similar among the patients with standard Ph, ACA, and vPh at diagnosis; these data are summarized in Table 1. The variant translocations involved chromosomes 1 (n Z 3), 6 (n Z 1), 7 (n Z 1), 15 (n Z 1), and 19 (n Z 1), and two patients had an insertional Ph and der(22), respectively. The most common ACAs were on chromosomes Y (n Z 3), 1 (n Z 2), 8 (n Z 2), and 10 (n Z 2). Two patients had different chromosomal abnormalities in two different clones.

Comparison with response and survival by CA at diagnosis Response and survival in the patients with vPh (n Z 9), ACA(s) to the Ph (n Z 13), and the patients with standard Ph (n Z 259) are shown in Table 2. Of the nine patients with vPh at diagnosis, six patients achieved CCyR without vPh during continuing IM therapy. Patient 6, with a t(1;9;22), experienced a loss of CCyR with the original vPh reappearing after 41 months of IM therapy and achieved subsequent CCyR with radotinib (RAD) therapy. Patient 7, with a t(9;22;15), experienced the disappearance of all of the original vPh, but 70% of the classical Ph chromosomes were identified at 4 months of IM treatment. This patient now maintains CCyR with ponatinib (PT) therapy. Patient 8, with an insertional vPh, progressed to AP after 5 months of IM therapy and died. Of the 13 patients with ACA at diagnosis, 10 patients’ abnormalities were lost with CCyR after IM therapy, but two patients, patient 20 with a t(10;12) and patient 21 with two different clones, including del(11) and del(18), lost CCyR after 39 and 25 months of IM treatment, respectively. Patient 20 lost the response with the classical Ph appearance, and patient 21 lost response with the emergence of a new clone, which had new CAs (add(11), add(17), marker chromosome) in addition to the original CAs (del(11), del(18)). Three patients (patients 18 and 19 with eY, and patient 17 with a t(1;22)) failed to respond to IM treatment. Patient 17, with the t(1;22), showed no CCyR after 3 months of IM therapy. Patient 18 showed minimal CCyR with the persistence of the Table 1

original clone after 13 months of IM therapy, and patient 19 lost complete hematologic response (CHR) with the detection of the classical Ph after 18 months of IM therapy. Overall, for a median follow-up of 78.6 months (range, 1.4e126.1 mo), for patients with a vPh, the 5-year OS, EFS, and FFS were 77.8%, 75.0%, and 53.3%, respectively. OS, EFS, and FFS were 100%, 66.3%, and 52.1%, respectively, for patients with ACA; 96.0%, 91.3%, and 83.7%, respectively, for patients with a standard Ph (Figure 1). Patients with vPh had a lower OS (P Z 0.003) and FFS (P Z 0.010) than those patients with the standard Ph. Further, ACA at diagnosis was associated with lower EFS (P Z 0.015) and FFS (P Z 0.016).

The characteristics of CA on therapy In eight patients, including one of the 22 (4.5%) patients with ACA or vPh at diagnosis and seven of the 259 (2.7%) patients with standard Ph at diagnosis, newly developed OCA in Phcells were found after a median of 12.4 months (range, 6.1e102.4 mo) from the start of IM therapy. The median age of these patients was 58 years (range, 35e68 y), and this group included four men and four women. Eight different OCAs from eight patients were detected during IM treatment, including þ8 (n Z 3), -Y (n Z 2), t(16;20) (n Z 1), t(1;21) (n Z 1), add(2) (n Z 1), del(3) (n Z 1), -7 (n Z 1), and a marker chromosome (n Z 2). The median percentage of metaphase cells involved was 35% (range, 15e100%) at the time of detection. Figure 2 shows the sequential changes in CAs detected at diagnosis and OCA development on IM therapy.

Response and survival in patients with OCA in Ph- cells on IM therapy The response and survival in each of the patients with OCA on IM therapy are shown in Table 3. Among the eight patients with OCA on IM therapy, six patients continued on IM therapy with optimal response (five in MMR and one in CCyR but no MMR) for a median of 6.8 years. Two patients

Patient demographic and clinical characteristics At diagnosis

Parameters

On therapy

Total (n Z 281) No CA (n Z 259) vPh (n Z 9 ) ACA (n Z 13) P

Age of patient, y, median (range) 41 Sex of patient, n (%) Male 161 Female 120 Interval from diagnosis to 0.6 treatment, mo, median (range) Sokal score, n (%) Low 103 Intermediate 83 High 41 Unknown 54 Imatinib duration, mo, median (range) 53.1

b

(18e77)

41 (18e77)

(57) (43) (0e6)

150 (58) 109 (42) 0.5 (0e6)

(37) (29) (15) (19) (1e125)

97 75 38 49 55.2

(37) (29) (15) (19) (1e125)

51 (32e70)

37 (23e67)

3 (33) 6 (67) 1.2 (0e4)

8 (62) 5 (38) 0.4 (0e6)

2 5 0 2 41.1

(22) 4 (56) 3 (0) 3 (22) 3 (4e104) 36.6

a

0.487 0.325

0.272

OCA (n Z 8b) 58 (35e68) 4 (50) 4 (50) 0.3 (0e4)

0.280 (31) 1 (23) 1 (23) 2 (23) 4 (10e104) 0.417 74.1

(12) (12) (26) (50) (12e104)

a Values for categorical variables were analyzed using a c2 or Fisher exact test to compare the characteristics of three CA groups at diagnosis. One-way ANOVA were utilized to compare the continuous variables. b One patient who had a vPh at diagnosis, loss of original CA and developed new marker chromosomes on treatment was included in both groups.

566

Table 2

Response and survival of patients with ACA or vPh at diagnosis

Patient Age, y Sex CA at diagnosis Variant 1 2 3 4 5 6 7 8

At 12 mo IM therapy IM duration, Cytogenetic Molecular Cause of IM response response discontinuation/mutation mo

Abbreviations: NA, not available; PON, ponatinib.

Survival, mo

complete minor complete complete complete complete NA NA

NA no MMR no MMR MMR no MMR MMR NA NA

none none none none none Loss of CCyR at 41 mo/no mutation minimal CyR at 4 mo/no mutation Progression at 5 mo/no mutation

MMR/IM MMR/IM CCyR/IM MMR/IM CCyR/IM CCyR/RAD CCyR/PON AP/-

Alive/63.0 Alive/104.4 Alive/79.6 Alive/103.8 Alive/37.5 Alive/60.6 Alive/60.0 Death/5.2

complete

MMR

Death

MMR/-

Death/19.8

complete

MR4.5

none

CCyR/IM

Alive/80.0

complete

MR4.5

none

MR4.5/IM

Alive/103.8

complete complete

no MMR no MMR

none none

MMR/IM MMR/IM

Alive/82.9 Alive/48.7

complete complete

MMR MMR

none none

MMR/IM MR4.5/IM

Alive/31.9 Alive/31.4

complete

no MMR

no MMR at 48 mo

CCyR/NIL

Alive/55.2

NA minimal partial complete

NA no MMR no MMR no MMR

no CyR at 3 mo/M244V minimal CyR at 13 mo/no mutation Loss of CHR at 18 mo/F359I Loss of CCyR at 39 mo/no mutation

CCyR/BOS CCyR/RAD MMR/DAS MR4.5/DAS

Alive/101.5 Alive/31.7 Alive/50.0 Alive/66.0

complete

NA

Loss of CCyR at 25 mo/M244V, G250E CCyR/BOS Alive/104.4

NA

NA

HSCT on CCyR

MMR/-

Alive/104.3

S.-E. Lee et al.

Philadelphia chromosome 53 F 46,XX,t(9;22;19)(q34;q11.2;p13)[20] 63 39 F 46,XX,t(7;9;22)(p22;q34;q11.2)[20] 104 40 F 46,XX,t(1;9;22)(p13;q34;q11.2)[20] 80 35 F 46,XX,der(22)t(9;22)(q34;q11.2)[20] 104 51 M 46,XY,t(6;9;22)(q23;q34;q11.2)[20] 38 32 M 46,XY,t(1;9;22)(p36.1;q34;q11.2)[20] 41 51 F 46,XX,t(9;22;15)(q34;q11.2;q22)[20] 4 70 F 46,XX[20].ish ins(9;22)(q34;q11.2) 5 (ABL1þ,BCRþ;BCRþ)[20] 9 54 M 46,XY,t(1;9;22)(p32;q34;q11.2)[20] 20 Additional chromosome abnormalities 10 67 F 46,XX,del(7)(q22q22),t(9;22)(q34;q11.2), 80 del(11)(q23)[19]/46,XX[1] 11 42 M 46,XY, t(9;22)(q34;q11.2)[17]/91,XXY,-Y, 104 t(9;22)(q34;q11.2)x2[2]/50,Y,-X,þY,-3,-6,þ7,-8, t(9;22)(q34;q11.2),þ12,þ13,þ17,þ17,þ21, þder(22)t(9;22)[1] 12 37 F 46,XX,t(9;22)(q34;q11.2)[15]/47,idem,þ8[5] 83 13 34 M 46,XY,?hsr(10)(q22),der(22)t(9;22)(q34;q11.2) 48 [10]/46,XY[10] 14 29 M 46,XY,t(9;22)(q34;q11.2),-21,þ22,þmar[4]/46,XY[6] 32 15 66 F 46,XX,der(9)t(9;22)(q34;q11.2)del(9)(q22q34), 31 der(22)t(9;22)[20] 16 25 M 46,XY,der(1)ins(9;1)(q34;q25q32)t(9;22)(q34;q11.2), 49 der(9)ins(9;1),der(22)t(9;22)[20] 17 37 M 46,XY,t(1;22)(q23;q13),t(9;22)(q34;q11.2)[20] 37 18 66 M 45,X,-Y,t(9;22)(q34;q11.2)[20] 14 19 36 M 46,XY,t(9;22)(q34;q11.2)[1]/45,idem,-Y[4] 20 20 23 F 46,XX,t(9;22)(q34;q11.2)[15]/46,idem, 39 t(10;12)(p11.2;p13)[5] 21 45 M 46,XY,del(11)(q23),t(9;22)(q34;q11.2) [6]/46,idem, 34 del(18)(q21)[2]/46,XY[12] 22 40 M 46,XY,t(9;22)(q34;q11.2)[2]/47,idem,þ8[18] 10

Current response/ treatment

Clonal CA in CML after imatinib

567

Figure 1 Survival outcomes in patients subdivided according to CA at diagnosis. KaplaneMeier estimates of (A) OS, (B) EFS, and (C) FFS in patients subdivided according to CA at diagnosis.

were resistant to IM treatment (patient 29 progressed to BC, and patient 28 exhibited suboptimal cytogenetic response) at the time of OCA development (one patient had add(2), del(3), -7, and þmar, one patient had þ8). Specifically, trisomy 8 was detected in patients 26, 27, and 28 while in suboptimal cytogenetic response (PCyR) after 12 months of IM treatment, and this abnormality was observed after the improvement in their response to MMR (n Z 2) with

continuing IM treatment and after MR4.5 (n Z 1) with a change to bosutinib (BOS) therapy. Patients 23 and 25 developed the loss of Y in CCyR after 85 and 6 months, respectively, from the beginning of IM therapy and had remained with CCyR and MR4.5 response at the last followup, 6.8 and 82.3 months after the detection of the abnormalities. Patient 24 developed a t(1;21) in CCyR after 10.7 months of IM therapy, and this abnormality disappeared

Figure 2 Sequential changes in CAs detected at diagnosis and OCAs that developed during IM therapy. Red indicates the patients with vPh; blue, patients with ACA; green, patients with OCA. The numbers indicate the percentages of vPh or ACA or OCA cells/ standard Ph cells/normal cells. *Two patients exhibited different chromosomal abnormalities in two different clones. yPatient 4 had a vPh at diagnosis, lost the original CA, and developed new marker chromosomes while in MMR after 102 months of IM treatment. ┃Time of cytogenetic test. C IM discontinuation. A Loss of CHR or CCyR.

568

Table 3

Response and survival of patients with OCA during IM treatment

Patient Age, y Sex OCA clones

After 12 mo IM therapy Cause of IM Current IM Time to Cytogenetic response discontinuation/ response/ Prior duration, clonal mutation treatment therapy mo evolution, mo (molecular)

Survival, mo

OCAs in Ph-negative cells 23 61 M 46,XY,t(16;20)(q24;p11.2)[2]/45,X,-Y[3]/46,XY[25]

IM

91

85.1

complete

none

CCyR/IM

Alive/91.1

a

24

57

M

46,XY,t(1;21)(p13;q22)[3]/46,XY[9]

IM

76

10.7

complete

none

MMR/IM

Alive/75.9

25

68

M

45,X,-Y[3]/46,XY[17]

IM

88

6.1

complete

none

MR4.5/IM

Alive/88.4

26

46

M

46,XY,t(9;22)(q34;q11.2)[6]/47,XY,þ8[9]/46,XY[5]

IM

65

12.5

partial

none

MMR/IM

Alive/65.2

27

46

F

46,XX,t(9;22)(q34;q11.2)[7]/47,XX,þ8[4]/46,XX[9]

IM

72

12.3

partial

none

MMR/IM

Alive/72.3

28

59

F

46,XX,t(9;22)(q34;q11.2)[2]/47,XX,þ8[17]/46,XX[1] IM

12

12.3

partial

29

62

F

45,XX,add(2)(q33),del(3)(q26.2),-7[7]/46, idem,þmar[13]

IM

19

19.3

complete

Suboptimal MR4.5/BOS Alive/86.0 response at 12 mo Progression at BC/Death/26.5 19 mo/F317F

4a

35

F

47,XX,þmar[20]

IM

104

102.4

complete (MMR)

none

MMR/IM

Alive/103.8

Patient 4 who had a vPh at diagnosis, loss of original CA, and developed new marker chromosomes on treatment was included in both groups.

S.-E. Lee et al.

Clonal CA in CML after imatinib spontaneously before a subsequent analysis. In patient 4, who had a vPh at diagnosis, the original CA disappeared and new marker chromosomes developed during MMR after 102.4 months of IM treatment; this response was maintained. In patient 29, abnormalities including add(2), del(3), -7, and a marker chromosome occurred in CCyR after 19.3 months of IM therapy; this patient progressed to BC at seven months after OCA detection and died. In summary, patient 29 with a monosomy 7, including multiple OCAs progressed to BC, and the remaining patients maintained long-term optimal responses with IM or BOS. There was no evidence of dysplasia by BM aspirate in these patients at the time of or after the detection of OCA. Patients 23 and 24 showed normal cellularity without dysplastic features, and a BM biopsy specimen from patient 29 showed hypocellularity without dysplastic features. In this study, of the 128 patients who continued on IM therapy with optimal response for more than five years, 86 were available for cytogenetic data collection. Among them, five different OCAs from six patients (7%) were detected.

Multivariate analyses as a time-dependent covariate To confirm the effect of OCA on outcomes, we performed univariate and multivariate analyses. Univariate analyses revealed that patients with ACA or vPh at diagnosis and high Sokal risk group were significantly associated with poor EFS and FFS (Supplementary Table S1). Multivariate analyses (Supplementary Table S2), including OCA on IM therapy as a time-dependent covariate in addition to the previously mentioned time-fixed variables, revealed that the presence of ACA or vPh at diagnosis was associated with lower EFS (P Z 0.021) and FFS (P Z 0.008), whereas the development of OCA in Ph- cells during IM therapy affected neither EFS nor FFS. Additionally, patients who were considered at intermediate Sokal risk had lower EFS (P Z 0.047) compared with that of patients at the low Sokal risk, high Sokal risk was an independent prognostic factor of lower EFS and FFS, and treatment was associated with OS.

Discussion Our study analyzed the prognostic impact of clonal CAs in a cohort of 281 consecutive patients with newly diagnosed CP CML who received IM as a first-line treatment. This study has significance because it evaluates the response to IM therapy in a homogeneous population and correlates the clinical course of the disease with much longer-term serial cytogenetic data than that reported in previous studies (6,10,13e15), in addition to considering the time and type of CA. In this study, the patients with ACA or vPh at diagnosis showed poor EFS and FFS. However, there was no association with OS because of the most likely impacts of secondline therapies such as dasatinib (DAS), nilotinib (NIL), RAD, and BOS. During IM therapy, seven of the 259 patients with standard Ph at diagnosis developed OCA, and one of the 22 patients with ACA or vPh at diagnosis developed new marker chromosomes with the disappearance of the original CA. Overall, there was no difference in the occurrence of OCA according to the absence or presence of CAs except

569 standard Ph (2.7% vs.4.5%, P Z 0.484), which demonstrates that the presence of ACA or vPh at diagnosis does not promote the development of OCA during IM therapy. Although ACA at diagnosis has traditionally been considered one of the major criteria for AP (18), the clinical impact of CA at diagnosis has been controversial (8,19). Hsiao et al. found that among 84 CML patients studied (12 in AP, 72 in CP), six patients had vPh and nine patients had ACA; they showed that ACA at diagnosis was associated with poor outcome over a mean follow-up of 40.8 months, whereas vPh had no impact on survival (5). In contrast, another study of 72 CML patients (21 in AP, 50 in CP with previous treatment with interferon, one in relapse after stem cell transplant [SCT]), including 49 patients with only the Ph chromosome and 23 patients with one or more ACAs at the start of IM treatment, showed that there was no difference in OS between patients with and without ACA (7). In our previous study, similar findings were observed in patients that were treated with IM after relapse subsequent to allografting (20). However, these results should be interpreted with caution because of the variety of disease phases, the effects of various prior treatments, and the short duration of follow-up. Patients who are heavily exposed to long-term interferon therapy and chemotherapy may have more CA (14,21). Recently, the German IV study evaluated the impact of ACA at diagnosis in a large patient cohort (9). However, the patients in this study were treated with IM alone or in combination with interferon or cytarabine, and the cytogenetic data were assessed only at diagnosis, which may make it difficult to estimate the clinical impact of ACAs on response to IM therapy and cytogenetic dynamics during IM therapy. Thus, in this study, we selected consecutive new CP patients, and found that the patients with ACA at diagnosis had lower EFS and FFS and that the patients with vPh also had inferior OS and FFS. These results are consistent with the previous report from Luatti et al., which showed ACA constitute an adverse prognostic factor in CML patients treated with IM as a frontline therapy (22). During IM treatment, the occurrence of CA in Phþ cells has been considered to represent treatment failure, which is supported by our two cases that developed CA in Phþ cells on IM therapy; one patient with del(7p) progressed to BC after 6.1 months of IM treatment, and the other patient lost CCyR with the occurrence of a t(1;17) and del(7q) after 51.8 months of IM treatment. Currently, this patient is alive and received allogeneic stem cell transplantation after the failure of second-line NIL therapy. Although the appearance of isolated OCA in Ph- cells is not considered to require a change in treatment strategy, several studies of OCA showed that the prognostic significance of this development is controversial, and analyses of the results of long-term treatment with IM are needed (13e15). In our study, none of the eight patients with OCA acquired any additional CA during the follow-up, except monosomy 7 in a patient with multiple OCAs, indicating that OCAs are unlikely to cause additional clonal changes related to resistance. These results were inconsistent with those of Jabbour et al., who showed an inferior PFS in patients with OCA (13). The differences in the frequency of the OCA in metaphases and the follow-up time after OCA detection may be possible explanations for this discrepancy. In addition, although our patients with OCA showed no evidence of dysplasia by BM aspirate, the

570 assessment of the clinical significance of OCA should also be correlated with pathological evaluation of possible therapy-related myelodysplastic syndrome (MDS). To date, although a few studies have shown that trisomy 8 in Ph- cells is transient and not related to MDS, the clinical significance of this feature is still unknown (23,24). Notably, our analyses showed that trisomy 8 in Ph- cells occurred in three patients with PCyR at 12 months and were detected in two patients after achieving CCyR and MMR with continuing IM treatment. One patient changed to BOS after suboptimal cytogenetic response and achieved MR4.5 even with trisomy 8 in Ph- cells. These results imply that IM treatment selectively eradicates the clones that express the Ph chromosome, and, thus, Ph- cells with trisomy 8 were identified. This observation might support the findings of Cannella et al., which suggested that the Ph chromosome developed as a secondary event in a fragile clone (25). Although recent studies have reported that major route abnormalities such as trisomy 8 at diagnosis were related to a worse outcome (9,22), we found two patients with trisomy 8 at diagnosis who did not show poor outcomes; one patient continued on IM therapy with optimal response for 83 months, and another patient who underwent SCT on CCyR after 10 months of IM therapy is alive without relapse at 94.3 months from SCT. However, it is difficult to give an exact answer to the clinical impact of the individual type of ACA at diagnosis because of the small number of patients with ACA. Lippert et al. evaluated the prognostic value of the loss of the Y chromosome and found that it is significantly associated with delayed cytogenetic and molecular responses, lower EFS, and shorter OS (26). In our study, two of three patients with eY at diagnosis failed to respond to IM treatment, whereas two patients that developed eY in Ph- cells during IM therapy continued on IM therapy with CCyR and MR4.5, respectively. In summary, this study showed that vPh at diagnosis was associated with poor OS and FFS, and ACA had affected EFS and FFS with IM therapy. In addition, there are distinct differences in the prognostic impact of CA in Phþ cells and in Ph- cells during IM therapy. These results may be considered to identify patients that require an increased dose of IM or a change in therapeutic strategy. Further studies about the influence of individual CA in Ph- cells with larger cohort sizes are needed.

Acknowledgments This study was supported by a grant from the National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea (1020400).

Supplementary data Supplementary data related to this article can be found online at http://dx.doi.org/10.1016/j.cancergen.2012.09.003.

References 1. Anastasi J, Feng J, Le Beau MM, et al. The relationship between secondary chromosomal abnormalities and blast transformation in chronic myelogenous leukemia. Leukemia 1995;9:628e633.

S.-E. Lee et al. 2. Swolin B, Weinfeld A, Westin J, et al. Karyotypic evolution in Ph-positive chronic myeloid leukemia in relation to management and disease progression. Cancer Genet Cytogenet 1985;18: 65e79. 3. Zaccaria A, Testoni N, Valenti AM, et al. Chromosome abnormalities additional to the Philadelphia chromosome at the diagnosis of chronic myelogenous leukemia: pathogenetic and prognostic implications. Cancer Genet Cytogenet 2010;199:76e80. 4. Mitelman F. The cytogenetic scenario of chronic myeloid leukemia. Leuk Lymphoma 1993;11:11e15. 5. Hsiao HH, Liu YC, Tsai HJ, et al. Additional chromosome abnormalities in chronic myeloid leukemia. Kaohsiung J Med Sci 2011;27:49e54. 6. Richebourg S, Eclache V, Perot C, et al. Mechanisms of genesis of variant translocation in chronic myeloid leukemia are not correlated with ABL1 or BCR deletion status or response to imatinib therapy. Cancer Genet Cytogenet 2008;182:95e102. 7. Holzerova M, Faber E, Veselovska J, et al. Imatinib mesylate efficacy in 72 previously treated Philadelphia-positive chronic myeloid leukemia patients with and without additional chromosomal changes: single-center results. Cancer Genet Cytogenet 2009;191:1e9. 8. Cortes JE, Talpaz M, Giles F, et al. Prognostic significance of cytogenetic clonal evolution in patients with chronic myelogenous leukemia on imatinib mesylate therapy. Blood 2003;101: 3794e3800. 9. Fabarius A, Leitner A, Hochhaus A, et al. Impact of additional cytogenetic aberrations at diagnosis on prognosis of CML: longterm observation of 1151 patients from the randomized CML Study IV. Blood 2011;118:6760e6768. 10. Marin D, Milojkovic D, Olavarria E, et al. European LeukemiaNet criteria for failure or suboptimal response reliably identify patients with CML in early chronic phase treated with imatinib whose eventual outcome is poor. Blood 2008;112:4437e4444. 11. Gorusu M, Benn P, Li Z, et al. On the genesis and prognosis of variant translocations in chronic myeloid leukemia. Cancer Genet Cytogenet 2007;173:97e106. 12. Marzocchi G, Castagnetti F, Luatti S, et al. Variant Philadelphia translocations: molecular-cytogenetic characterization and prognostic influence on frontline imatinib therapy, a GIMEMA Working Party on CML analysis. Blood 2011;117:6793e6800. 13. Jabbour E, Kantarjian HM, Abruzzo LV, et al. Chromosomal abnormalities in Philadelphia chromosome negative metaphases appearing during imatinib mesylate therapy in patients with newly diagnosed chronic myeloid leukemia in chronic phase. Blood 2007;110:2991e2995. 14. Deininger MW, Cortes J, Paquette R, et al. The prognosis for patients with chronic myeloid leukemia who have clonal cytogenetic abnormalities in philadelphia chromosome-negative cells. Cancer 2007;110:1509e1519. 15. Kovitz C, Kantarjian H, Garcia-Manero G, et al. Myelodysplastic syndromes and acute leukemia developing after imatinib mesylate therapy for chronic myeloid leukemia. Blood 2006;108: 2811e2813. 16. Baccarani M, Cortes J, Pane F, et al. Chronic myeloid leukemia: an update of concepts and management recommendations of European LeukemiaNet. J Clin Oncol 2009;27:6041e6051. 17. Shaffer LG, Slovak ML, Campbell LJ, editors. ISCN 2009: an International System for Human Cytogenetic Nomenclature. Basel: Karger; 2009. 18. Lahaye T, Riehm B, Berger U, et al. Response and resistance in 300 patients with BCR-ABL-positive leukemias treated with imatinib in a single center: a 4.5-year follow-up. Cancer 2005; 103:1659e1669. 19. O’Dwyer ME, Mauro MJ, Kurilik G, et al. The impact of clonal evolution on response to imatinib mesylate (STI571) in accelerated phase CML. Blood 2002;100:1628e1633. 20. Kim YJ, Kim DW, Lee S, et al. Cytogenetic clonal evolution alone in CML relapse post-transplantation does not adversely affect

Clonal CA in CML after imatinib response to imatinib mesylate treatment. Bone Marrow Transplant 2004;33:237e242. 21. Fayad L, Kantarjian H, O’Brien S, et al. Emergence of new clonal abnormalities following interferon-alpha induced complete cytogenetic response in patients with chronic myeloid leukemia: report of three cases. Leukemia 1997;11:767e771. 22. Luatti S, Castagnetti F, Marzocchi G, et al. Additional chromosomal abnormalities in Philadelphia-positive clone: adverse prognostic influence on frontline imatinib therapy: a GIMEMA Working Party on CML analysis. Blood 2012;120:761e767. 23. Gozzetti A, Tozzuoli D, Crupi R, et al. Emergence of Ph negative clones in chronic myeloid leukemia (CML) patients in complete cytogenetic remission after therapy with imatinib mesylate (STI). Eur J Haematol 2003;71:313e314.

571 24. Kim M, Lee S, Jung CK, et al. Transient trisomy 8 abnormality in Philadelphia-negative cells during imatinib mesylate treatment of chronic myelogenous leukemia. Int J Lab Hematol 2008;30: 508e512. 25. Cannella L, Loglisci G, Nanni M, et al. Trisomy 8 in Philadelphia chromosome negative cell preceding the evolution of a Philadelphia chromosome positive clone with the same additional change during imatinib treatment: revisiting the role of genetic instability in chronic myeloid leukemia. Leuk Lymphoma 2012; 53:497e498. 26. Lippert E, Etienne G, Mozziconacci MJ, et al. Loss of the Y chromosome in Philadelphia-positive cells predicts a poor response of chronic myeloid leukemia patients to imatinib mesylate therapy. Haematologica 2010;95:1604e1607.