Introduction

Chronic myelogenous leukemia (CML) constitutes a model disease for biology, novel treatment approaches and therapy surveillance of neoplastic disorders. More than 95% of cases are characterized by a typical genetic aberration, the translocation t(9;22)(q34;q11), with its molecular consequences – the BCR-ABL fusion gene, mRNA, and protein.^1,2 This biological background provided the basis for the development of novel methods of therapy surveillance and their clinical translation: banding cytogenetics in the 1970s,³ Southern blot analysis in the 1980s,^4,5,6 and reverse transcriptase-polymerase chain reaction (RT-PCR) in the 1990s.^7,8,9 Owing to the limited value of qualitative PCR for therapy monitoring, quantitative PCR assays were developed.^10,11,12 Most recently, the advent of real-time methods simplified the procedure and permitted clinical routine use.^13,14,15,16

The introduction of the selective tyrosine kinase inhibitor imatinib (STI571, Glivec^®) in the treatment of CML rapidly revealed the need for a standardized sensitive method to measure response. Complete cytogenetic remission (CCR) was achieved in 41–45% of chronic-phase patients pretreated with interferon alpha (IFN )^17,18 and 74% of newly diagnosed patients¹⁹ within the first 18 months on treatment. However, due to the limited sensitivity of cytogenetics, levels of residual disease may vary by several orders of magnitude even in these good responders.

We investigated molecular response of patients enrolled into the International Randomized Study of Interferon and STI571 (IRIS study) in 17 German centers. Residual BCR-ABL transcript levels in CML patients in early chronic phase on drug treatment were monitored, independent of the cytogenetic response status. The aim of this study was to determine (i) quantitative levels of residual disease in comparison to cytogenetic response; (ii) the optimum reference gene to calculate response levels; (iii) the dynamics of response during imatinib vs IFN/cytarabine (Ara-C) therapy and after crossover from IFN/Ara-C to imatinib; (iv) molecular response patterns within prognostic risk groups and according to BCR-ABL transcript types; (v) cytogenetic response in relation to pretherapeutic transcript levels; and (vi) risk of relapse according to best molecular response.

Patients, materials and methods

Patients and therapy

A total of 139 patients (79 male, 60 female; median age 51 years, range 20–71) with newly diagnosed BCR-ABL-positive CML in chronic phase were studied prospectively. Patients were recruited into the multicenter phase III trial 0106 (IRIS study) between June 2000 and January 2001¹⁹ in Germany and comprise 12.6% of all patients treated in this study (n=1106). Primary aim of the randomized study was to compare the efficacy of imatinib with that of IFN combined with low-dose Ara-C with regard to hematological and cytogenetic response and progression-free survival.

A molecular substudy was set up in Germany on the basis of a protocol amendment, aiming to determine the dynamics of BCR-ABL mRNA expression in the newly diagnosed CML patients treated with imatinib or IFN/Ara-C. After written informed consent, patients were randomized to primary therapy with imatinib (n=69) or IFN/Ara-C (n=70). Risk distribution according to the prognostic scores was not different between treatment groups: Hasford score:²⁰ 43 vs 40% low-risk patients; Sokal score:²¹ 43 vs 42% low-risk patients in the imatinib and IFN/Ara-C groups, respectively. Patients in the imatinib group received 400 mg orally daily. Patients assigned to IFN/Ara-C received gradually escalating doses of IFN with a target dose of 5 million U/m² daily. After reaching the maximal tolerated dose of IFN, Ara-C was added s.c. at a dose of 20 mg/m² (maximal daily dose 40 mg) daily for 10 days every month. The concurrent administration of hydroxyurea in either treatment group was permitted during the first 6 months.¹⁹ According to the protocol and with permission of the study management committee, 36 patients discontinued IFN/Ara-C treatment on the basis of lack of response or severe intolerance and switched to imatinib therapy.

Cytogenetic analysis

Cytogenetic analyses were performed on BM aspirates (n=803, 0–10 per patient, median 6) according to standard protocols. Metaphases from direct and/or short-term (24 h) cultures were examined after Giemsa banding. Cytogenetic data were obtained in the imatinib arm, n=436; the IFN/Ara-C arm, n=248; and after crossover from IFN/Ara-C to imatinib, n=119. Cytogenetic response was categorized as complete (CCR, 0% Ph+ metaphases), partial (PR, 1–34%), minor (MR, 35–95%), or nonresponse (NR, 96–100%). For comparison with molecular data, only results derived from at least 20 metaphases were considered.²²

Sample collection and processing

Peripheral blood (PB) samples for molecular analysis were centrally collected prior to therapy, and after 1, 2, and 3 months and at three-monthly intervals thereafter and were received either locally or by mail and spent 24–48 h in transit. Previous studies demonstrated that PB can be used instead of BM monitoring the tumor load in chronic-phase CML.^11,23,24 Total white blood cells were isolated from 20 ml of PB by hypotonic red cell lysis, washed and transferred into a guanidinium isothiocyanate-containing buffer. RNA was extracted by commercially available kits (RNeasy, Qiagen, Hilden, Germany) or by CsCl gradient centrifugation. cDNA synthesis was performed using random hexamer primers and MMLV reverse transcriptase (Invitrogen, Karlsruhe, Germany). Conditions for RNA extraction and reverse transcription have been described.²⁵ cDNAs were stored at -20°C.

Multiplex PCR

Prior to therapy, the leading BCR-ABL transcript was determined in 135 cases by multiplex PCR using three BCR- and one ABL primers as described.²⁶ cDNAs derived from K562 (b3a2 BCR-ABL) and BV173 (b2a2 BCR-ABL) cell lines were amplified in parallel and served as controls. Rare BCR-ABL fusion variants were confirmed by direct sequencing.

Quantitative real-time RT-PCR

In all samples, quantitative real-time polymerase chain reaction (RQ-PCR) for BCR-ABL mRNA transcripts was performed (n=1117, 0–16 per patient, median 9), independent of the cytogenetic response status. Samples were analyzed to monitor imatinib (n=569), or IFN/Ara-C therapy (n=394), and after crossover from IFN/Ara-C to imatinib (n=154).

Real-time PCR was carried out using the LightCycler technology (Roche Diagnostics, Mannheim, Germany). Conditions for amplification, hybridization, and fluorescence detection have been described.¹⁴

The number of total ABL and glucose-6-phosphate dehydrogenase (G6PD) transcripts were determined as internal standards in all samples in order to adjust for different qualities of RNA and cDNA and to exclude negative results derived from samples of poor quality. The final results were calculated as the ratios BCR-ABL/ABL and BCR-ABL/G6PD and expressed in percent.

Qualitative nested and quantitative competitive PCR

Since dilutions of CP CML cells in normal PB cells revealed a higher sensitivity of nested PCR (10^-6) vs real-time PCR (10^-5),¹⁴ samples negative by real-time PCR were examined by qualitative 'nested' PCR for BCR-ABL transcripts.²⁵ BCR-ABL was considered 'undetectable' in case of three negative PCR tests to control for sampling effects²⁷ and an adequate number of ABL transcripts documenting the good quality of the sample. The lab-specific cutoff point for appropriate sample quality was 5 10⁴ ABL transcripts/5 l cDNA.²³

In samples positive by nested PCR and negative by real-time PCR, a quantitative competitive PCR titration assay was performed¹⁰ using a semilogarithmic dilution series of 10¹–10^4.5 competitor molecules added to the same volume of patients' cDNA. The equivalence point between different competitor concentrations was determined by densitometric analysis.^11,12

Mutation analysis

Patients with confirmed relapse after CCR were investigated for mutations of the BCR-ABL tyrosine kinase domain by PCR amplification and direct sequencing as described previously.²⁸

Statistical methods

PCR results were expressed as the ratios between BCR-ABL and total ABL and between BCR-ABL and G6PD transcripts, expressed as percentage. Correlation coefficients and P-values were calculated according to Spearman's rank test. Comparison of molecular results and cytogenetic response groups were performed using the ² test, and the nonparametric Kruskal–Wallis and Mann–Whitney tests. Overall cytogenetic and molecular responses were calculated on the basis of all patients recruited into the study, response at specific time points refers to the number of samples analyzed at this time points. All tests were calculated using the GraphPad Prism software (GraphPad Software, Inc., San Diego, CA, USA).

Results

Cytogenetic analysis

Cytogenetic response was determined in 67/69 patients after a median interval of imatinib therapy of 24 months (range 1–30). CCR was achieved in 60 (87.0%), PR in three (4.3%), and MR in four patients (5.8%). In the IFN/Ara-C group, cytogenetic response was determined after a median observation time of 18 months (range 1–30) and revealed CCR in 10 (14.3%), PR in 13 (18.6%), MR in 19 (27.1%), and no response in 19 patients (27.1%). After a median time of IFN/Ara-C treatment of 7.5 months (range 1.6–27), 36 patients crossed over to imatinib therapy and were followed over a median observation time of 13 months (range 0–25). Cytogenetic response to second-line imatinib therapy was CCR in 28 (77.7%), PR in two (5.6%), MR in three (8.3%), and NR in two patients (5.6%) (Table 1).

Table 1 - Best cytogenetic response during therapy with imatinib (n=69), IFN/Ara-C (n=70), and after crossover from IFN/Ara-C to imatinib (n=36).

Full table

Qualitative determination of the BCR-ABL fusion

Qualitative multiplex RT-PCR was performed prior to therapy in 135 patients and showed 56 patients with b2a2 (41.5%), 55 with b3a2 (40.7%), 22 with b3a2 and b2a2 (16.3%), and one each with b3a3 (0.7%) and e19a2 (0.7%) BCR-ABL transcripts. There was no significant difference in the distribution of the transcript types between the treatment arms.

Quantification of BCR-ABL compared to cytogenetics

Ratios BCR-ABL/ABL derived from RQ-PCR of PB samples were compared to contemporaneously performed cytogenetic analyses from BM metaphases. In imatinib-treated patients, median ratios BCR-ABL/ABL prior to therapy were 51% (range 1.1–210, n=98), after CCR 0.067% (range 0–5.7%, n=85), in PR 1.4% (range 0.18–11%, n=5), in MR 27% (range 6.1–69%, n=7), and in NR 42% (range 38–45%, n=2) (Figure 1).

Figure 1.

Comparison of ratios BCR-ABL/ABL determined by RQ-PCR and cytogenetic response groups. The differences between the cytogenetic response groups are significant (P<0.0001). (CR – complete, PR – partial, MR – minor, NR – no cytogenetic remission).

Full figure and legend (23K)

Cytogenetic response to imatinib (complete, partial, or minor/none) was compared with molecular response by introducing cutoff points in the BCR-ABL/ABL and BCR-ABL/G6PD ratios. For the ratio BCR-ABL/ABL, the optimum cutoff points described after IFN therapy¹¹ of 2 and 14% were used (ie, a ratio <2% was compared with CCR, between 2 and 14% with PR, and >14% with minor and nonresponders). For the ratio BCR-ABL/G6PD, optimum cutoff points of 0.13 and 1% were empirically established, the ratio BCR-ABL/G6PD <0.13% was correlated to CCR, 0.13–1% to PR, and >1% to minor and NR. For the ratio BCR-ABL/ABL, 173/236 samples (73%), for the ratio BCR-ABL/G6PD, 163/236 samples (69%) were concordant between the two methods (² test P<0.0001, Table 2).

Table 2 - Comparison of cytogenetic and molecular response status in imatinib-treated patients on the basis of simultaneously analyzed BM and PB sample pairs (n=236). The concordance between the two methods was 173/236 samples (73%, ratio BCR-ABL/ABL) and 163/236 samples (69%, ratio BCR-ABL/G6PD).

Full table (15K)

Quantification of control gene mRNA expression

In all samples, total ABL and G6PD transcripts were quantified as internal controls. Ratios BCR-ABL/ABL and BCR-ABL/G6PD correlated well (r=0.82, P<0.0001). However, in samples with low level of residual disease (BCR-ABL/ABL ratio <1.0%), the median ratio G6PD/ABL in the individual sample was 5.8 with a wide range of 0.03–200.

Prediction of response to imatinib

Ratios BCR-ABL/G6PD prior to imatinib therapy were compared from patients who consecutively reached CCR within 12 months (n=45, median 2.7%, range 0.076–27) with patients lacking CCR (n=17, median 1.0%, range 0.30–8.8, NS). The response to imatinib was independent from the level of BCR-ABL transcripts prior to therapy. Further, there was no difference between initial BCR-ABL/G6PD ratios in patients with CCR on IFN/Ara-C therapy (n=5, median 3.9%, range 0.34–7.6) compared to patients without CCR (n=22, median 2.9%, range 0.71–80, NS).

After 2 months of imatinib therapy, there was a trend to predict CCR within 12 months using both the ratio BCR-ABL/ABL (P=0.084) and BCR-ABL/G6PD (P=0.19), but after 3 months of treatment, CCR within the first year could be predicted using the ratio BCR-ABL/ABL (P=0.0026) or BCR-ABL/G6PD (P=0.0074) (Table 3). Empirically derived statistical cutoff points for best prediction of CCR after 12 months were a ratio BCR-ABL/ABL of 10% with a positive predictive value of 71% and a negative predictive value of 82% and a reduction of the ratio BCR-ABL/G6PD of 0.3 log after 3 months with a positive predictive value of 76% and a negative predictive value of 80%, respectively.

Table 3 - Prediction of CCR to imatinib therapy within 12 months using molecular results prior to therapy, and after 2 and 3 months.

Full table

The treatment response was not different according to the type of transcripts as determined by the CCR rate and the proportion of patients achieving ratios BCR-ABL/ABL <0.1%. One patient with a rare b3a3 BCR-ABL transcript achieved PCR negativity after crossover to imatinib therapy. Another patient with the rare e19a2 BCR-ABL transcript reached only MR to imatinib therapy. Three patients with undetectable BCR-ABL after imatinib therapy carried b3a2 BCR-ABL transcripts.

Sensitivity of the methods

On imatinib therapy, 60 patients were negative by cytogenetics, 23 by RQ-PCR, and three by nested PCR. On IFN/Ara-C, 10 patients were negative by cytogenetics, all of them revealed residual disease by RQ-PCR. After crossover from IFN/Ara-C to imatinib, 29 patients were negative by cytogenetics, six by RQ-PCR, and one patient by nested PCR. Thus, BCR-ABL was undetectable in four patients (Table 4).

Table 4 - Levels of disease elimination expressed as the number of patients tested positive with cytogenetics, real-time RT-PCR, and nested RT-PCR.

Full table

In 39 samples with negative RQ-PCR results, nested competitive PCR was employed to quantify low levels of residual disease. In all, 28/33 samples from patients on primary imatinib therapy, 4/4 samples from patients on IFN/Ara-C therapy, and 2/2 samples from patients after crossover from IFN/Ara-C therapy were positive by competitive PCR and revealed a ratio BCR-ABL/ABL which was used for further calculations. Nested competitive PCR showed a median of 34 (range 0–380) BCR-ABL transcripts/2.5 l cDNA and 80 000 (range 14 000–320 000) ABL transcripts/2.5 l cDNA, resulting in a median ratio BCR-ABL/ABL of 0.046% (range 0–0.76).

Response dynamics during therapy

For all samples analyzed, there was a gradual reduction of the residual disease within 24 months of imatinib therapy (median 3.1 log), IFN/Ara-C (median 1.6 log), and imatinib after IFN/Ara-C (median 2.3 log) considering the ratios BCR-ABL/ABL (Figure 2). Using the independent control gene G6PD, the ratios BCR-ABL/G6PD showed a higher degree of heterogeneity at each time point as compared to the standardization to total ABL transcripts.

Figure 2.

Relative change of the ratios BCR-ABL/ABL after first CCR in patients on primary imatinib therapy (a); primary IFN/Ara-C therapy (b), and after crossover from IFN-Ara-C to imatinib (c). There is a gradual decrease of the median level of residual disease in all groups of patients.

Full figure and legend (78K)

Figure 3 illustrates the molecular response levels according to therapy and treatment interval. In imatinib-treated patients, a reduction of the ratio BCR-ABL/ABL by at least three orders of magnitude was achieved in 33/69 patients (47.8%), a ratio BCR-ABL/ABL <0.1% in 39/69 patients (56.5%) (Figure 3a). In contrast, a >3-log reduction was achieved in 3/70 (4.3%) of IFN/Ara-C-treated patients with a ratio BCR-ABL/ABL <0.1% in 4/70 (5.7%) patients (Figure 3b). Second-line imatinib therapy showed a similar pattern as primary imatinib and led to a >3-log reduction of the ratio BCR-ABL/ABL in 7/36 patients (19.4%) and to a ratio BCR-ABL/ABL <0.1% in 10/36 patients (27.8%).

Figure 3.

Histograms illustrating the proportion of patients reaching different levels of residual disease within 21 months after start of therapy with imatinib (a) and IFN/Ara-C (b). The columns represent the best response within the total observation time.

Full figure and legend (239K)

Molecular response according to prognostic risk at diagnosis

Molecular response to imatinib was not different between risk groups according to the Hasford score (low risk, n=20, median ratio BCR-ABL/ABL 0.11%, range 0.0019–11%; intermediate risk, n=24, median ratio BCR-ABL/ABL 0.036%, range 0–3.8%; high risk, n=2, ratios BCR-ABL/ABL 0.013 and 0.24%, NS), or the Sokal score (low risk, n=19, median ratio BCR-ABL/ABL 0.067%, range 0.0022–16%; intermediate risk, n=17, median ratio BCR-ABL/ABL 0.15%, range 0–6.1%; high risk, n=7; median ratio BCR-ABL/ABL 0.022%, range 0.0090–2.8% (NS, Figure 4). Further, median molecular response was not different at all time points after start of imatinib therapy according to Hasford-risk groups.

Figure 4.

Best molecular response to imatinib was not different in the three risk groups according to the Hasford (a) and Sokal (b) scores.

Full figure and legend (31K)

Minimal residual disease levels after achieving CCR

Ratios BCR-ABL/ABL at the time of first CCR were not different in patients on imatinib and IFN/Ara-C therapy (median 0.51 vs 0.80%, NS). There was a gradual decrease of residual disease levels after CCR in imatinib-treated patients with median BCR-ABL/ABL ratios of 0.26% after 3, 0.16% after 6, 0.093% after 9, 0.048% after 12, 0.058% after 15, 0.033% after 18, 0.026% after 21, 0.030% after 24, and 0.020% after 27 months (P<0.0001, Figure 5). Referring to the ratio BCR-ABL/ABL at first CCR, levels of residual disease in imatinib-treated patients declined faster within the first 12 months ( of the median ratio BCR-ABL/ABL –0.77 log, n=48) as compared to patients on IFN/Ara-C ( of the median ratio –0.47 log, n=9, P=0.025). The decline of residual disease levels in CCR after crossover from IFN/Ara-C to imatinib (-0.74 log, n=14) was faster than that in CCR on IFN/Ara-C (but NS) and comparable to the decline on primary imatinib therapy (NS).

Figure 5.

Dynamics of BCR-ABL/ABL ratios in imatinib-treated patients after achieving CCR. There was a gradual decrease of the median BCR-ABL transcripts.

Full figure and legend (28K)

Undetectable BCR-ABL

In four of 139 patients studied (2.9%), residual disease levels declined below the detection limit of all methods used. In total, seven samples were tested negative. The number of ABL transcripts/5 l cDNA in these samples was median 1.4 10⁵ (range 5.5 10⁴–2 10⁵). Three patients were on primary imatinib therapy, one on imatinib after IFN/Ara-C intolerance. CCR was achieved after 3 (n=3) or 6 months (n=1) of imatinib therapy: Patient 1 (male, 62 years old) reached PCR negativity after 18 months of imatinib therapy. Patient 2 (male, 68 years old) was tested PCR negative after 21 months, but a follow up test after 29 months revealed a low level of BCR-ABL transcripts. In patient 3 (male, 51 years old), PCR negativity was reached after 24 months of imatinib therapy and confirmed at month 30. Patient 4 (male, 49 years old) was primarily treated with IFN/Ara-C and crossed over to imatinib due to severe adverse events and loss of hematologic remission after 7 months. Undetectable BCR-ABL levels were reached after 8 months of imatinib therapy (month 15 of treatment in study), this status was confirmed at months 21 and 28 (Figure 6).

Figure 6.

Gradual decrease of the ratio BCR-ABL/ABL in whom BCR-ABL transcripts became undetectable. Patients 1–3 were primarily treated with imatinib, patient 4 was intolerant to IFN/Ara-C and crossed over to imatinib at month 7.

Full figure and legend (27K)

Cytogenetic relapse

Of 60 patients, 16 (26.7%) experienced reoccurrence of Ph+ metaphases (median 8%, range 3–100) after a median of 6 months (range 3–21) following CCR on primary treatment with imatinib. In three of those 16 patients, cytogenetic relapse was sustained while 13 patients regained CCR during further follow up continuing imatinib therapy. One patient (male, 64 years old, high risk according to the Hasford score) who had reached CCR after 9 months of imatinib therapy gradually relapsed with 3% Ph+ metaphases after 12 months to 79% Ph+ metaphases after 18 months accompanied by clonal evolution (tetrasomy 8, +17, +Y). An escalation of the imatinib dosage to 800 mg/day resulted in an improvement of cytogenetic response to 7% Ph positivity. The second patient (male, 43 years old) with CCR after 12 months showed 14% Ph+ metaphases after 15 months and 100% Ph+ metaphases after 24 months. The third patient (male, 22 years old, low risk) achieved CCR after 3 months and showed 5% Ph+ metaphases at month 18 and 45% at month 24. At the time of CCR, best response in patients who subsequently relapsed was 0.24% (n=3, range 0.15–3.1) as compared to 0.029% in patients with continuous remission (n=52, range 0–1.1, P=0.028). Five additional patients lack cytogenetic follow-up results after first CCR. In all, 3/28 patients with best ratios BCR-ABL/ABL >0.1% have relapsed in contrast to 0/37 patients with a best ratio BCR-ABL/ABL <0.1%. One of the 28 patients with CCR after crossover to imatinib has relapsed. CCR was achieved after 5 months of imatinib therapy with a best ratio BCR-ABL/ABL of 2.5%. He showed 36% Ph+ at month 14 and 96% Ph+ at month 18. Three of four patients with confirmed cytogenetic relapse after CCR were investigated for mutations of the BCR-ABL tyrosine kinase domain, but all revealed the wild-type sequence between exons 4 and 7 of the ABL part of the fusion gene.

Discussion

Based on the high hematologic and cytogenetic response rate with imatinib monotherapy, molecular response has been accepted as a major end point for new studies. Several prospective clinical trials are conducted to investigate imatinib in different dosages or in combination with other antileukemic therapies in order to optimize the way to use imatinib in the treatment of CML.²⁹ The comparison of molecular responses after novel treatment strategies requires a standardized approach to measure and calculate residual disease levels and an idea as to which responses are expected with current imatinib monotherapy as reference level.

We therefore analyzed molecular responses in 139 patients treated in 17 German institutions within the IRIS study. In contrast to the international PCR studies, analyses were performed on all patients in study, independent of their cytogenetic response and even after crossover to the alternative therapy. Blood processing and PCR analyses were performed in a central laboratory according to standardized protocols. The patients investigated here represent a cohort (12.6%) of the IRIS Study population (n=1106) with identical median age (51 years), but with a less favorable risk profile than in the total study (Hasford score: low risk 40.7 vs 45.1%; Sokal score: low risk 43.8 vs 50.3%). Response to imatinib was determined after a median treatment interval of 24 months as compared to 19 months of the total study.¹⁹ As a consequence, CCR rate was higher after imatinib therapy with 87.0 vs 73.8%, after IFN/Ara-C with 14.3 vs 8.5%, and after crossover from IFN/Ara-C to imatinib with 77.7 vs 39.6% in this study.

Molecular responses were measured by RQ-PCR. At the time of first CCR, ratios BCR-ABL/ABL were not different within the treatment groups. However, the speed of response before and after CCR was significantly different. In the first year after CCR, median levels of minimal residual disease decreased by 0.77 log on imatinib vs 0.47 log on IFN/Ara-C. This is comparable to data reported by other groups (Branford et al, Blood 2002; 100 (Suppl. 1): 96a, abstract) and confirms data from a previous study demonstrating a reduction by 0.53 log after 1 year, and 1.36 log after 5 years of IFN therapy.¹² Although the most significant decline of residual disease on imatinib happened within the first 3 months of therapy (median reduction by 1.0 log), levels continued to decrease up to 24 months (median reduction by 3.1 log) (Figure 2a).

Quantitative PCR data on imatinib therapy parallel cytogenetic response. However, there is a significant overlap of individual molecular data between the cytogenetic response groups, partly caused by the statistical uncertainty with conventional cytogenetics.³⁰ However, the optimum cutoff points for molecular response are identical to those derived from IFN therapy:¹¹ CCR matches ratios BCR-ABL/ABL of <2%, PR between 2 and 14%, and MR and NR >14%. New cutoff points were calculated for ratios BCR-ABL/G6PD with <0.13% for CCR, 0.13–1% for PR, and >1% for MR and NR.

Several parameters were investigated for predictability of cytogenetic and molecular response. BCR-ABL transcript type, prognostic risk groups, and pretherapeutic levels of BCR-ABL transcripts were not predictive for cytogenetic and molecular response on imatinib therapy within the treatment interval analyzed. In contrast, an association of clinical risk status and cytogenetic response has been observed in a study combining imatinib with pegylated IFN 2b.³¹

The earliest time point demonstrating a significant difference of transcript levels between patients reaching CCR and not after 1 year of therapy was 3 months. This result supplements data of chronic-phase CML patients intolerant or resistant to IFN. In this group of patients, CCR at 6 months was predicted by RQ-PCR after 2 months of imatinib treatment.³²

Within the observation time of median 24 months, three patients relapsed after CCR. Thus, there is limited information on the predictive value of residual disease levels for long-term response. However, all patients who consecutively relapsed had at that time CCR ratios BCR-ABL/ABL of >0.1%. Ratios BCR-ABL/ABL <0.1% were achieved in 39/69 patients (56.5%) on first-line and in 10/69 patients (27.8%) on second-line imatinib therapy. Longer follow up will certainly reduce the level of residual disease that can be used as a landmark associated with long-term remission. After a median treatment duration of 4 years with IFN, ratios BCR-ABL/ABL of <0.045% have been associated with long-term event-free survival and low relapse rates.^12,29 More recently, we found in a longitudinal study on chronic-phase CML patients treated with imatinib after IFN failure, ratios BCR-ABL/ABL <0.1% being associated with continuous CCR.²³

The question of how to express response data (absolute vs relative results) is still conflicting. It has been suggested for minimal residual disease monitoring in general that kinetics of response to therapy could be more important than absolute data at a single time point.³³ Therefore, as a relative measure, the proportion of patients achieving a reduction of residual disease by at least three orders of magnitude was determined. A >3-log reduction of BCR-ABL transcript levels was achieved after 12 months in 27.6% of patients on imatinib therapy. The rate of >3-log reduction was lower in this unselected group of patients than in PCR data from the total study, concentrating on CCR patients only (39%, Hughes et al, Blood 2002, 100 (Suppl. 1): 93a–94a, abstract). However, the overall rate of a >3-log reduction after a median observation time of 2 years was 56.5 vs 5.7% on IFN/Ara-C. The disadvantage of using relative changes is the requirement of pretherapeutic data as point of reference.

Using strict criteria for the definition of undetectable BCR-ABL, PCR negativity was observed in only three patients on first-line and one patient on second-line imatinib therapy. Criteria included a negative RQ-PCR, three negative nested PCRs, and a level of ABL transcripts >5 10⁴/5 l cDNA. However, as with cytogenetic remission, the correlation of RT-PCR negativity with durability of remission, although likely, remains to be proven.³⁴ Residual disease may represent a clone with additional molecular abnormalities that mediate resistance. Another possibility is that quiescent stem cells may be insensitive to imatinib.³⁵ Inhibition of BCR-ABL tyrosine kinase activity by imatinib does not eliminate malignant primitive progenitors in CML patients.³⁶ Alternatively, different stages of the hematopoietic stem cell development could be involved in the initial pathogenesis of CML. If the stem cells have differential sensitivity to imatinib, this may explain the variability in cytogenetic and molecular responses.³⁷

The data reported by several groups on BCR-ABL monitoring after imatinib therapy in CML are heterogeneous. The differences might be explained by the methodological variability between centers. The differences relate to the rate of patients with undetectable BCR-ABL, early predictability of response, and the use of PB vs BM.^{23,32,38,39,40}

Until recently, the use of quantitative RT-PCR data for therapeutic decisions in CML was not accepted on a general basis.⁴¹ However, the clinical impact of quantitative PCR data has been clearly demonstrated by systematic studies performed in specialized laboratories after allogeneic stem cell transplantation^10,42 and after IFN therapy.^12,43 The clinical need for molecular end points has become even more compelling with the introduction of imatinib in the therapy of CML. Again, certain degrees of molecular response may become useful surrogate markers for predicting outcome and will gain power with longer follow-up times.

We conclude that (i) because of the high cytogenetic complete response rate observed with imatinib, the appropriate measure for response in CML has now become molecular remission using PCR detection of residual BCR-ABL transcripts; (ii) treatment with imatinib in newly diagnosed CML patients is associated with a rapid decrease of BCR-ABL transcript levels; (iii) nested PCR may reveal residual BCR-ABL transcripts in samples which are negative by real-time PCR; (iv) BCR-ABL transcript levels correlate with cytogenetic response, and (v) imatinib is superior to IFN/Ara-C in terms of speed and degree of molecular responses, but residual disease is rarely eliminated.

References

1. Sawyers CL. Chronic myeloid leukemia. N Engl J Med 1999; 340: 1330–1340. | Article | PubMed | ISI | ChemPort |
2. Thijsen S, Schuurhuis G, van Oostveen J, Ossenkoppele G. Chronic myeloid leukemia from basics to bedside. Leukemia 1999; 13: 1646–1674. | Article | PubMed |
3. Rowley JD. Letter: a new consistent chromosomal abnormality in chronic myelogenous leukaemia identified by quinacrine fluorescence and Giemsa staining. Nature 1973; 243: 290–293. | Article | PubMed | ISI | ChemPort |
4. Groffen J, Stephenson JR, Heisterkamp N, de Klein A, Bartram CR, Grosveld G. Philadelphia chromosomal breakpoints are clustered within a limited region, bcr, on chromosome 22. Cell 1984; 36: 93–99. | Article | PubMed | ISI | ChemPort |
5. Grossman A, Mathew A, O'Connell MP, Tiso P, Distenfeld A, Benn P. Multiple restriction enzyme digests are required to rule out polymorphism in the molecular diagnosis of chronic myeloid leukemia. Leukemia 1990; 4: 63–64. | PubMed |
6. Reiter A, Skladny H, Hochhaus A, Seifarth W, Heimpel H, Bartram CR et al. Molecular response of CML patients treated with interferon-alpha monitored by quantitative Southern blot analysis. German chronic myeloid leukaemia (CML) Study Group. Br J Haematol 1997; 97: 86–93. | Article | PubMed | ISI | ChemPort |
7. Morgan GJ, Hughes T, Janssen JW, Gow J, Guo AP, Goldman JM et al. Polymerase chain reaction for detection of residual leukaemia. Lancet 1989; 1: 928–929. | Article | PubMed | ISI | ChemPort |
8. Hughes TP, Morgan GJ, Martiat P, Goldman JM. Detection of residual leukemia after bone marrow transplant for chronic myeloid leukemia: role of polymerase chain reaction in predicting relapse. Blood 1991; 77: 874–878. | PubMed | ChemPort |
9. Lion T, Henn T, Gaiger A, Kalhs P, Gadner H. Early detection of relapse after bone marrow transplantation in patients with chronic myelogenous leukaemia. Lancet 1993; 341: 275–276. | Article | PubMed |
10. Cross NC, Feng L, Chase A, Bungey J, Hughes TP, Goldman JM. Competitive polymerase chain reaction to estimate the number of BCR-ABL transcripts in chronic myeloid leukemia patients after bone marrow transplantation. Blood 1993; 82: 1929–1936. | PubMed | ISI | ChemPort |
11. Hochhaus A, Lin F, Reiter A, Skladny H, Mason PJ, van Rhee F et al. Quantification of residual disease in chronic myelogenous leukemia patients on interferon-alpha therapy by competitive polymerase chain reaction. Blood 1996; 87: 1549–1555. | PubMed | ISI | ChemPort |
12. Hochhaus A, Reiter A, Saussele S, Reichert A, Emig M, Kaeda J et al. Molecular heterogeneity in complete cytogenetic responders after interferon-alpha therapy for chronic myelogenous leukemia: low levels of minimal residual disease are associated with continuing remission. German CML Study Group and the UK MRC CML Study Group. Blood 2000; 95: 62–66. | PubMed | ISI | ChemPort |
13. Mensink E, van de Locht A, Schattenberg A, Linders E, Schaap N, Geurts van Kessel A et al. Quantitation of minimal residual disease in Philadelphia chromosome positive chronic myeloid leukaemia patients using real-time quantitative RT-PCR. Br J Haematol 1998; 102: 768–774. | Article | PubMed | ISI | ChemPort |
14. Emig M, Saussele S, Wittor H, Weisser A, Reiter A, Willer A et al. Accurate and rapid analysis of residual disease in patients with CML using specific fluorescent hybridization probes for real time quantitative RT-PCR. Leukemia 1999; 13: 1825–1832. | Article | PubMed | ISI | ChemPort |
15. Preudhomme C, Revillion F, Merlat A, Hornez L, Roumier C, Duflos-Grardel N et al. Detection of BCR-ABL transcripts in chronic myeloid leukemia (CML) using a 'real time' quantitative RT-PCR assay. Leukemia 1999; 13: 957–964. | Article | PubMed | ISI | ChemPort |
16. Hochhaus A, Weisser A, La Rosée P, Emig M, Müller MC, Saussele S et al. Detection and quantification of residual disease in chronic myelogenous leukemia. Leukemia 2000; 14: 998–1005. | Article | PubMed | ISI | ChemPort |
17. Kantarjian H, Sawyers C, Hochhaus A, Guilhot F, Schiffer C, Gambacorti-Passerini C et al. Hematologic and cytogenetic responses to imatinib mesylate in chronic myelogenous leukemia. N Engl J Med 2002; 346: 645–652. | Article | PubMed | ISI | ChemPort |
18. Kantarjian HM, Talpaz M, O'Brien S, Smith TL, Giles FJ, Faderl S et al. Imatinib mesylate for Philadelphia chromosome-positive, chronic-phase myeloid leukemia after failure of interferon-alpha: follow-up results. Clin Cancer Res 2002; 8: 2177–2187. | PubMed | ISI | ChemPort |
19. O'Brien SG, Guilhot F, Larson RA, Gathmann I, Baccarani M, Cervantes F et al. Imatinib compared with interferon and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med 2003; 348: 994–1004. | Article | PubMed | ISI | ChemPort |
20. Hasford J, Pfirrmann M, Hehlmann R, Allan NC, Baccarani M, Kluin-Nelemans JC et al. A new prognostic score for survival of patients with chronic myeloid leukemia treated with interferon alfa. Writing Committee for the Collaborative CML Prognostic Factors Project Group. J Natl Cancer Inst 1998; 90: 850–858. | Article | PubMed | ChemPort |
21. Sokal JE, Cox EB, Baccarani M, Tura S, Gomez GA, Robertson JE et al. Prognostic discrimination in 'good-risk' chronic granulocytic leukemia. Blood 1984; 63: 789–799. | PubMed | ISI | ChemPort |
22. Schoch C, Schnittger S, Bursch S, Gerstner D, Hochhaus A, Berger U et al. Comparison of chromosome banding analysis, interphase- and hypermetaphase-FISH, qualitative and quantitative PCR for diagnosis and for follow-up in chronic myeloid leukemia: a study on 350 cases. Leukemia 2002; 16: 53–59. | Article | PubMed | ISI | ChemPort |
23. Paschka P, Müller MC, Merx K, Kreil S, Schoch C, Lahaye T et al. Molecular monitoring of response to imatinib (Glivec) in CML patients pretreated with interferon alpha. Low levels of residual disease are associated with continuous remission. Leukemia 2003; 17: 1687–1694. | Article | PubMed | ISI | ChemPort |
24. Lin F, Goldman JM, Cross NC. A comparison of the sensitivity of blood and bone marrow for the detection of minimal residual disease in chronic myeloid leukaemia. Br J Haematol 1994; 86: 683–685. | PubMed | ISI |
25. Cross NC, Feng L, Bungey J, Goldman JM. Minimal residual disease after bone marrow transplant for chronic myeloid leukaemia detected by the polymerase chain reaction. Leuk Lymphoma 1993; 11 (Suppl 1): 39–43. | PubMed | ISI |
26. Cross NC, Melo JV, Feng L, Goldman JM. An optimized multiplex polymerase chain reaction (PCR) for detection of BCR-ABL fusion mRNAs in haematological disorders. Leukemia 1994; 8: 186–189. | PubMed | ISI | ChemPort |
27. Melo JV, Yan XH, Diamond J, Lin F, Cross NC, Goldman JM. Reverse transcription/polymerase chain reaction (RT/PCR) amplification of very small numbers of transcripts: the risk in misinterpreting negative results. Leukemia 1996; 10: 1217–1221. | PubMed |
28. Hochhaus A, Kreil S, Corbin AS, La Rosée P, Müller MC, Lahaye T et al. Molecular and chromosomal mechanisms of resistance to imatinib (STI571) therapy. Leukemia 2002; 16: 2190–2196. | Article | PubMed | ISI | ChemPort |
29. Schiffer CA, Hehlmann R, Larson R. Perspectives on the treatment of chronic phase and advanced phase CML and Philadelphia chromosome positive ALL (1). Leukemia 2003; 17: 691–699. | Article | PubMed | ISI | ChemPort |
30. Hook EB. Exclusion of chromosomal mosaicism: tables of 90%, 95% and 99% confidence limits and comments on use. Am J Hum Genet 1977; 29: 94–97. | PubMed | ISI | ChemPort |
31. Rosti G, Trabacchi E, Bassi S, Bonifazi F, de Vivo A, Martinelli G et al. Risk and early cytogenetic response to imatinib and interferon in chronic myeloid leukemia. Haematologica 2003; 88: 256–259. | PubMed |
32. Merx K, Müller MC, Kreil S, Lahaye T, Paschka P, Schoch C et al. Early reduction of BCR-ABL mRNA transcript levels predicts cytogenetic response in chronic phase CML patients treated with imatinib after failure of interferon alpha. Leukemia 2002; 16: 1579–1583. | Article | PubMed | ISI | ChemPort |
33. van der Velden V, Hochhaus A, Cazzaniga G, Szczepanski T, Gabert JA, van Dongen JJ. Detection of minimal residual disease in hematologic malignancies by real-time quantitative PCR: principles, approaches, and laboratory aspects. Leukemia 2003; 17: 1013–1034. | Article | PubMed | ISI | ChemPort |
34. Deininger MW, O'Brien SG, Ford JM, Druker BJ. Practical management of patients with chronic myeloid leukemia receiving imatinib. J Clin Oncol 2003; 21: 1637–1647. | Article | PubMed | ChemPort |
35. Graham SM, Jorgensen HG, Allan E, Pearson C, Alcorn MJ, Richmond L et al. Primitive, quiescent, Philadelphia-positive stem cells from patients with chronic myeloid leukemia are insensitive to STI571 in vitro. Blood 2002; 99: 319–325. | Article | PubMed | ISI | ChemPort |
36. Bhatia R, Holtz M, Niu N, Gray R, Snyder DS, Sawyers CL et al. Persistence of malignant hematopoietic progenitors in chronic myelogenous leukemia patients in complete cytogenetic remission following imatinib mesylate treatment. Blood 2003; 101: 4701–4707. | Article | PubMed | ISI | ChemPort |
37. Druker BJ. Perspectives on the development of a molecularly targeted agent. Cancer Cell 2002; 1: 31–36. | Article | PubMed | ISI | ChemPort |
38. Kantarjian HM, Talpaz M, Cortes J, O'Brien S, Faderl S, Thomas D et al. Quantitative polymerase chain reaction monitoring of BCR-ABL during therapy with imatinib mesylate (STI571; Gleevec) in chronic-phase chronic myelogenous leukemia. Clin Cancer Res 2003; 9: 160–166. | PubMed | ISI | ChemPort |
39. Wang L, Pearson K, Ferguson JE, Clark RE. The early molecular response to imatinib predicts cytogenetic and clinical outcome in chronic myeloid leukaemia. Br J Haematol 2003; 120: 990–999. | Article | PubMed | ISI | ChemPort |
40. Stentoft J, Pallisgaard N, Kjeldsen E, Holm MS, Nielsen JL, Hokland P. Kinetics of BCR-ABL fusion transcript levels in chronic myeloid leukemia patients treated with STI571 measured by quantitative real-time polymerase chain reaction. Eur J Haematol 2001; 67: 302–308. | Article | PubMed | ISI | ChemPort |
41. Faderl S, Talpaz M, Kantarjian HM, Estrov Z. Should polymerase chain reaction analysis to detect minimal residual disease in patients with chronic myelogenous leukemia be used in clinical decision making? Blood 1999; 93: 2755–2759. | PubMed |
42. Lin F, van Rhee F, Goldman JM, Cross NC. Kinetics of increasing BCR-ABL transcript numbers in chronic myeloid leukemia patients who relapse after bone marrow transplantation. Blood 1996; 87: 4473–4478. | PubMed | ISI | ChemPort |
43. Kantarjian H, Talpaz M, O'Brien S, Giles F, Beth RM, White K et al. Prediction of initial cytogenetic response for subsequent major and complete cytogenetic response to imatinib mesylate therapy in patients with Philadelphia chromosome-positive chronic myelogenous leukemia. Cancer 2003; 97: 2225–2228. | Article | PubMed |

Acknowledgements

We thank the members of the German CML Study Group, coinvestigators, nursing and research staff for the excellent cooperation. The study was supported by Novartis Pharma GmbH, Nürnberg, Germany, the Forschungsfonds der Fakultät für Klinische Medizin Mannheim der Universität Heidelberg, Germany, and the Competence Network 'Acute and chronic leukemias', sponsored by the German Bundesministerium für Bildung und Forschung (Projektträger Gesundheitsforschung; DLR e.V.- 01 GI9980/6). We are grateful to Ms Insa Gathmann, Novartis Pharmaceuticals, Basel, Switzerland, for providing clinical and cytogenetic data.