A nomogram for predicting T315I-free survival in chronic phase chronic myeloid leukemia patients: a multicenter retrospective study

doi:10.21203/rs.3.rs-4943970/v1

Download PDF

Article

A nomogram for predicting T315I-free survival in chronic phase chronic myeloid leukemia patients: a multicenter retrospective study

https://doi.org/10.21203/rs.3.rs-4943970/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The T315I mutation poses a significant threat to patients with chronic phase chronic myeloid leukemia (CP-CML). This study aimed to establish a nomogram to predict the risk of T315I mutation in CP-CML patients. The training cohort included 1,466 patients from 24 hematology centers, and the validation cohort included 820 patients from an additional 20 centers. Peripheral blood blast (PBB), additional chromosomal abnormality (ACA), dasatinib use, non-EMR at 3 months, and BCR::ABL^IS > 1% at 6 months were identified as independent risk factors through multivariate Cox regression analysis. The performance of the nomogram was assessed via receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). The area under the ROC curve (AUC) values at 5, 10, and 15 years were 0.874, 0.925, and 0.930 for the training cohort, and 0.864, 0.814, and 0.803 for the validation cohort, respectively. The calibration curves for both cohorts were close to the ideal diagonal, and the decision curves indicated clinical net benefit. In conclusion, we developed a nomogram to predict the 5-year, 10-year, and 15-year T315I-free survival probabilities of CP-CML patients. This tool can aid clinicians in the early prediction and timely management of high-risk CP-CML patients with the T315I mutation.

Biological sciences/Computational biology and bioinformatics/Predictive medicine

Health sciences/Oncology/Cancer

Health sciences/Risk factors

chronic myeloid leukemia

T315I mutation

prediction nomogram

Chronic myeloid leukemia (CML) is a malignant clonal disorder originating from hematopoietic stem cells in the bone marrow and is characterized by the presence of the Philadelphia chromosome (Ph) and the BCR::ABL1 fusion gene¹. The advent of tyrosine kinase inhibitors (TKIs) has revolutionized the treatment of CML, with survival life expectancies approaching those of the general population². Studies have shown that the 5-year overall survival (OS) rate for chronic phase CML (CP-CML) patients receiving first-line TKI therapy can reach 95%³, and the 10-year OS rate can reach 88%⁴.

However, many patients develop resistance or intolerance after continuous TKI treatment, leading to a decreased survival probability. A common mechanism of resistance is the emergence of BCR::ABL1 kinase domain (KD) mutations. Over 100 types of amino acid alterations within the BCR::ABL1 KD gene family have been identified to date⁵. Among these, the T315I gatekeeper mutation has the highest incidence, accounting for approximately 15–20% of all acquired mutations⁶. As the most aggressive mutation with the poorest prognosis, the T315I mutation is most commonly associated with disease progression and relapse⁷. The T315I mutation confers high resistance to imatinib, dasatinib, nilotinib, and bosutinib by affecting the affinity of TKIs for the ATP binding site, posing a significant threat to the prognosis of CML patients⁸.

The median survival time for CP-CML patients with the T315I mutation, accelerated-phase CML (AP-CML) patients, and blast-phase CML (BP-CML) patients is 133 months, 31 months, and 6 months, respectively⁹. Considering the poor prognosis in AP and BP, clinicians are more inclined to test for BCR::ABL1 KD mutations in AP-CML or BP-CML patients. Additionally, owing to the longer follow-up periods for CP-CML patients, there is a clinical tendency to overlook timely monitoring of BCR::ABL1 KD mutations in CP-CML patients, leading to a lower detection rate of the T315I mutation during CP.

However, research indicates that more than half of T315I mutations occur in the CP-CML rather than in the AP-CML or BP-CMLpatients⁹. Interestingly, our preliminary study revealed that, the T315I mutation is associated with a poorer prognosis compared with other BCR::ABL1 KD mutations, such as E255K/V, Y253H, and G250E in CP-CML patients. However, in AP-CML and BP-CML patients, the differences in prognosis among patients with various BCR::ABL1 KD mutations were not statistically significant. Therefore, exploring the occurrence of the T315I mutation in CP-CML patients is highly significant. It is crucial to have a tool that identifies high-risk CP-CML patients prone to developing the T315I mutation, facilitating timely switches to TKIs effective against T315I mutations.

Currently, no clinical predictive model for the T315I mutation in CML patients has been developed. As simple statistical visualization tools, nomograms have been widely used in recent years to predict disease occurrence, development, prognosis, and survival¹⁰. This retrospective study aimed to identify risk factors for the T315I mutation in CP-CML patients and to establish a nomogram to predict the 5-year, 10-year, and 15-year T315I-free survival probability in these patients. This work could support clinicians in conducting early assessments, implementing proactive measures, and improving patient outcomes.

Study design and participants

This multicenter retrospective investigation was conducted for the CML Cooperation Group of Zhejiang Hematology in China. From January 2010, to May 2023, a total of 1,956 CML patients were treated and followed up at 24 hematology centers. After 83 AP-CML patients and 108 BP-CML patients were excluded, a total of 1,860 CP-CML patients were included in the study. Among them, 83 patients had incomplete records, 168 patients had a follow-up period of less than 12 months, 48 patients were diagnosed younger than 18 years, and 95 patients had BCR::ABL1 KD mutations other than T315I. Ultimately, 1,466 CP-CML patients were included in the training cohort. The external validation cohort comprised 820 patients from 20 hematologic oncology centers between January, 2010, and May, 2023. The inclusion criteria were as follows: (1) met the diagnostic criteria for CML¹¹; (2) were aged older than 18 years; and (3) had a follow-up duration exceeding 12 months. The exclusion criteria were (1) having experienced AP or BP, (2) missing critical clinical records, and (3) having other BCR::ABL1 KD mutations excluding T315I.

Ethics statement

The study protocol was approved by the Ethics Committee of the First Affiliated Hospital, School of Medicine, Zhejiang University Institutional Review Board, with a waiver for informed consent (2022, No469). All methods were carried out in accordance with relevant guidelines and regulations .

Data collection and definition

The data collected included the patient's age, sex, additional chromosomal abnormality (ACA), spleen size below the costal margin, white blood cell count (WBC), hemoglobin level (HB), platelet count (PLT), eosinophil count (EOS), basophil count (BAS), peripheral blood blast (PBB), time from CML diagnosis to the start of first TKI, choice of TKI, history of TKI discontinuation, number of TKI lines used, number of lines of dasatinib used, duration of dasatinib use, and molecular response at 3, 6, and 12 months.

ACAs were defined as the major route ACA + 8, +Ph, i[17q], + 19, +21, + 17, the minor route ACA 3q26.2, 11q23, − 7/7q−, and complex karyotypes^12,13. The choice of TKI was categorized into patients who had used dasatinib and those who had not used dasatinib. The number of TKI lines was classified into 1-line TKI use (no TKI switch) and ≥ 2-line TKI use (one or more TKI switches). Number of lines of dasatinib used was classified as none (no prior use of dasatinib), 1st -line (asatinib used as first-line therapy), and ≥ 2nd -line (dasatinib used as second-line or third-line therapy). The response definitions were as follows: (1) early molecular response (EMR): BCR::ABL^IS ≤ 10% at 3 months; (2) BCR::ABL^IS ≤ 1% at 6 months; and (3) major molecular response (MMR): BCR::ABL^IS ≤ 0.1% at 12 months.

Development and assessment of the nomogram

Univariate Cox analysis was used to identify clinical parameters significantly associated with the occurrence of T315I. Variables with p values < 0.05 were included in the multivariate Cox analysis. Independent risk factors were used to construct a nomogram to predict the 5-year, 10-year, and 15-year T315I-free survival probabilities in CP-CML patients. Each predictive factor is assigned a line according to its weight, corresponding to different scores. The total score is the sum of the points for all variables in the nomogram. The T315I-free survival rates at different time points were obtained on the basis of the total score.

The model was validated via the concordance index (C-index), receiver operating characteristic (ROC) curve, area under the ROC curve (AUC), calibration curve, and decision curve analysis (DCA). The C-index and AUC were used to evaluate the discriminative ability of the model. The accuracy of the model was tested via calibration curves, which compare the observed values with the actual values via 1000 bootstrap resamples. DCA was used to assess the clinical potential application value of the nomogram by quantifying the net benefit at different threshold probabilities¹⁴. Additionally, we classified patients into low-risk and high-risk groups on the basis of their nomogram scores. K‒M curves and log-rank tests were used to compare survival differences between the risk groups. The model's discriminative ability, accuracy, and clinical net benefit in the external validation cohort were validated via the same methods.

Statistical analysis

All eligible patients in the CML Cooperation Group of Zhejiang Hematology from January, 2010 to May, 2023, were included in this analysis, and no additional sample size calculations were performed. All the data were analyzed via IBM SPSS Statistics Version 28.0 and R software version 4.3.0 with the rms, time‒ROC, survival, ggplot2 and dcurves packages. Categorical covariates are reported as percentages and counts. Continuous variables are reported as medians and interquartile ranges (IQRs). For comparisons among these groups, the Pearson chi-square test was used for categorical factors, and the Mann‒Whitney U test was used for continuous variables. P < 0.05 was considered to indicate statistical significance.

Baseline characteristics

Out of 2,051 CML patients from 24 hematology centers, 585 patients were excluded. Ultimately, 1,466 patients were included in the training cohort (Fig. 1). A total of 898 (61.3%) patients were male, with a median age of 50 years (IQR 37–60 years). The median follow-up time was 3.6 years (range 1–24.9 years), and 30 patients (2.0%) developed the T315I mutation. The 5-year, 10-year, and 15-year T315I-free survival probabilities were 98.3% (95% CI 97.6%-99.1%), 97.1% (95% CI 95.6%-98.7%), and 91.3% (95% CI 86.1%-96.8%), respectively (Fig. S1). The external validation cohort included 820 patients who met the inclusion and exclusion criteria, with a median follow-up time of 3.7 years (range 1–29.5 years), and 23 patients (2.8%) who developed the T315I mutation. The 5-year, 10-year, and 15-year T315I-free survival probabilities were 97.9% (95% CI 96.5%-99.2%), 92.4% (95% CI 88.8%-96.1%), and 88.3% (95% CI 81.6%-95.5%), respectively (Fig. S1). There were no statistically significant differences between the training and validation cohorts across the 19 listed parameters, including age, sex, WBC, HB, PLT, EOS, BAS, PBB, spleen size below the costal margin, ACA, time from diagnosis to first TKI, TKI discontinuation, TKI lines, TKI choice, lines of dasatinib, duration of dasatinib, EMR at 3 months, BCR::ABL^IS ≤ 1% at 6 months, and MMR at 12 months (Table 1).

Table 1

Baseline characteristics of the training cohort and validation cohort.
Variables	Training cohort (n = 1466) M (P25, P75)/N (%)	External validation cohort (n = 820) M (P25, P75)/N (%)	P-value
Age (years)	50 (37, 60)	49 (38, 61)	0.466
Sex			0.351
Male	898 (61.3)	486 (59.3)
Female	568 (38.7)	334 (40.7)
White blood cell (×10ˆ9/L)	92.80 (38.5, 159.5)	93.00 (38.0, 159.5)	0.623
Hemoglobin (g/L)	111.00 (98.0, 127.0)	110.00 (98.0, 128.0)	0.509
Platelet (×10ˆ9/L)	419.50 (262.0, 591.0)	424.00 (261.0, 579.3)	0.71
Eosinophils (%)	2.25 (1.1, 3.0)	2.33 (1.2, 3.3)	0.216
Basophils (%)	4.00 (2.0, 5.3)	4.00 (1.8, 5.1)	0.591
Peripheral blood blast (%)	0.00 (0.0, 2.0)	0.00 (0.0, 2.0)	0.092
Spleen size below costal margin (cm)	3.50 (0.0, 6.0)	3.50 (0.0, 5.7)	0.974
ACA (%)			0.517
Yes	77 (5.3)	38 (4.6)
No	1389 (94.7)	782 (95.4)
Time from diagnosis to first TKI (months)	0.00 (0.0, 0.0)	0.00 (0.0, 0.0)	0.44
TKI discontinuation (%)			0.666
Yes	133 (9.1)	70 (8.5)
No	1333 (90.9)	750 (91.5)
TKI lines (%)			0.821
1-line	1122 (76.5)	631 (77.0)
≥2-line	344 (23.5)	189 (23.0)
TKI choice (%)			0.424
Dasatinib	168 (11.5)	85 (10.4)
Other	1298 (88.5)	735 (89.6)
Lines of dasatinib			0.627
None	1298 (88.5)	735 (89.6)
1st -line	27 (1.8)	16 (2.0)
≥ 2nd -line	141 (9.6)	69 (8.4)
Duration of dasatinib (months)	0.00 (0.0, 0.0)	0.00 (0.0, 0.0)	0.418
EMR at 3 months			0.51
Yes	1107 (75.5)	609 (74.3)
No	359 (24.5)	211 (25.7)
BCR::ABL^IS ≤ 1% at 6 months			0.651
Yes	1067 (72.8)	604 (73.7)
No	399 (27.2)	216 (26.3)
MMR at 12 months			0.668
Yes	1044 (71.2)	577 (70.4)
No	422 (28.8)	243 (29.6)
ACA = additional chromosomal abnormalities; EMR = early molecular response; MMR = major molecular response

Nomogram model development

According to the univariate Cox regression analysis, WBC (P = 0.019), HB (P = 0.017), PBB (P = 0.001), ACA (P = 0.011), TKI lines (P = 0.01), TKI choice (P < 0.0001), lines of dasatinib (P = 0.001), EMR at 3 months (P < 0.0001), and BCR::ABL^IS ≤ 1% at 6 months (P < 0.0001) differed significantly between the T315I mutation group and the non-T315I mutation group (Table 2). These variables were then included in the multivariable Cox regression analysis. PBB (P = 0.042, hazard ratio (HR) = 1.171, 95% confidence interval (CI) 1.006–1.364), ACA (P = 0.020, HR = 3.712, 95% CI 1.224–11.255), TKI choice of dasatinib (P = 0.011, HR = 7.769, 95% CI 1.594–37.868), non-EMR at 3 months (P = 0.033, HR = 5.692, 95% CI 1.146–28.262), and BCR::ABL^IS > 1% at 6 months (P = 0.005, HR = 22.794, 95% CI 2.529–205.450) were found to be independent risk factors for the occurrence of the T315I mutation in CP-CML patients (Table 2).

Table 2

General characteristics of the patients and multivariate Cox regression analyses for the training group.
Variables	T315I mutation (n = 30) M (P25, P75)/N (%)	Non-T315I mutation (n = 1436) M (P25, P75)/N (%)	P-value	HR 95% CI
Age (years)	44 (31, 51)	50 (37, 61)	0.177
Sex			0.674
Male	19 (63.3)	879 (61.2)
Female	11 (36.7)	557 (38.8)
White blood cell (×10ˆ9/L)	151.6 (57.6, 244.7)	92.8 (38.2, 159.5)	0.019
Hemoglobin (g/L)	108.0 (87.8, 118.0)	112.0 (98.0, 127.0)	0.017
Platelet (×10ˆ9/L)	400.0 (246.5, 540.0)	423.0 (263.0, 591.0)	0.937
Eosinophils (%)	2.0 (1.25, 3.4)	2.3 (1.1, 3.0)	0.574
Basophils (%)	3.9 (1.8, 5.9)	4.0 (2.0, 5.3)	0.411
Peripheral blood blast (%)	1.0 (0.0, 4.0)	0.0 (0.0, 1.8)	0.001	1.171 (1.006,1.364)
Spleen size below costal margin (cm)	4.7 (0.5, 7.0)	3.5 (0.0, 6.0)	0.306
ACA (%)			0.011	3.712 (1.224,11.255)
Yes	4 (13.3)	73 (5.1)
No	26 (86.7)	1363 (94.9)
Time from diagnosis to first TKI (months)	0.0 (0.0, 1.0)	0.0 (0.0, 0.0)	0.912
TKI discontinuation (%)			0.253
Yes	7 (23.3)	126 (8.8)
No	23 (76.7)	1310 (91.2)
TKI lines (%)			0.010
1-line	14 (46.7)	1108 (77.2)
≥2-line	16 (53.3)	328 (22.8)
TKI choice (%)			< 0.0001	3.128 (1.016,9.631)
Dasatinib	14 (46.7)	154 (10.7)
Other	16 (53.3)	1282 (89.3)
Lines of dasatinib
None	16 (53.3)	1282 (89.3)
1st -line	2 (6.7)	25 (1.7)	0.011
≥ 2nd -line	12 (40.0)	129 (9.0)	< 0.0001
Duration of dasatinib (months)	0.0 (0.0, 15.3)	0.0 (0.0, 0.0)	0.54
EMR at 3 months			< 0.0001	5.692 (1.146,28.262)
Yes	2 (6.7)	1105 (76.9)
No	28 (93.3)	331 (23.1)
BCR::ABL^IS ≤ 1% at 6 months			< 0.0001	22.794 (2.529,205.450)
Yes	1 (3.3)	1066 (74.2)
No	29 (96.7)	370 (25.8)
MMR at 12 months			0.995
Yes	0 (0.0)	1044 (72.7)
No	30 (100.0)	392 (27.3)
ACA = additional chromosomal abnormalities; EMR = early molecular response; MMR = major molecular response

By integrating the above five factors, a nomogram model was constructed to predict the T315I-free survival probability at different time points for CP-CML patients (Fig. 2). Points of each variable in the nomogram were shown in the Table S1 in the supplement. As time progresses, higher scores indicate a lower T315I-free survival probability. For example, if a CP-CML patient presents with ACA, has a PBB of 2%, has a history of using dasatinib, does not achieve EMR at 3 months, and has a BCR::ABL^IS > 1% at 6 months, the corresponding scores would be approximately 42 points, 11 points, 26 points, 53 points, and 100 points, respectively. The total score would be 232 points, corresponding to a 5-year T315I-free survival probability of 68.7%, a 10-year survival rate of 55.0%, and a 15-year survival rate of 10.0%.

Nomogram model validation

In the training cohort, the C-index was 0.894, and the AUC values for 5, 10, and 15 years were 0.874 (95% CI 0.798–0.950), 0.925 (95% CI 0.857–0.989), and 0.930 (95% CI 0.864–0.995), respectively, indicating the model's long-term discriminative ability (Fig. 3A). The calibration curve was close to the ideal diagonal line, reflecting good accuracy (Fig. 4A-C). The DCA curve demonstrated a clinical benefit for intervention in patients whose predicted risk was within the threshold range (Fig. 5A-C). The nomogram model was validated in the validation cohort via the same methods. Analysis revealed a C-index of 0.857, and the AUC values at 5, 10, and 15 years were 0.864 (95% CI 0.731–0.998), 0.814 (95% CI 0.668–0.959), and 0.803 (95% CI 0.624–0.981), respectively (Fig. 3B). The calibration curve in the validation cohort was also close to the diagonal (Fig. 4D-F), and the DCA curve demonstrated clinical net benefit within a certain range (Fig. 5D-F).

Clinical application of the nomogram

According to the ROC curve, the cutoff value for the total score was approximately 128 points. Patients with scores > 128 were defined as the high-risk group, whereas those with scores ≤ 128 were defined as the low-risk group. In both the training and validation cohorts, the Kaplan‒Meier curves revealed that patients in the high-risk group had a lower T315I-free survival probability than patients in the low-risk group did (Fig. 6A and B). In the overall population, the 5-year and 10-year T315I-free survival probabilities for the low-risk group were 99.6% (95% CI 99.3%-100.0%) and 98.8% (95% CI 97.6%-100.0%), respectively, whereas they were 92.9% (95% CI 90.1%-95.7%), 83.3% (95% CI 77.6%-89.5%), and 60.6% (95% CI 47.4%-77.4%), respectively, for the high-risk group.

The survival rate of CP-CML patients is close to that of the general population; however, CP-CML patients with concurrent T315I mutations have a poorer prognosis. In the imatinib era, the median OS of CP-CML patients with the T315I mutation was found to be 22.4 months¹⁵. In the ponatinib era, the 5-year survival rate for CP-CML patients with the T315I mutation was 70%⁹. Currently, treatment options for patients with the T315I mutation primarily include third-generation TKIs (such as ponatinib, asciminib, and olverembatinib) and allogeneic hematopoietic stem cell transplantation (HSCT).⁵ Research indicates that in CP-CML patients with the T315I mutation, ponatinib is superior to allogeneic HSCT¹⁶. Typically, the timing for detecting BCR::ABL1 KD mutations is based on increasing levels of BCR::ABL1 transcripts⁷, but low-level T315I mutations often occur before clinical detection¹⁷, Therefore, timely identification of patients with the T315I mutation can help physicians implement appropriate interventions to improve disease prognosis.

Several studies have explored the relevant factors and predictive models for molecular response, TKI resistance, and treatment failure in CML patients^18–20. However, a model for predicting the T315I mutation in CP-CML patients has yet to be developed. Here, we developed an innovative and meaningful nomogram model for evaluating the risk of T315I mutation in CP-CML patients. We included 1,466 eligible CP-CML patients from 24 hematology centers. The incidence of the T315I mutation was 2.0%, similar to the 4.2% incidence reported in a recent study⁹. From the training cohort of 1,466 patients, we identified five independent risk factors influencing the occurrence of the T315I mutation in CP-CML patients: PBB, ACA, choice of dasatinib as a TKI, non-EMR at 3 months, and BCR::ABL^IS > 1% at 6 months. We developed a nomogram model based on these factors to predict the 5-year, 10-year, and 15-year T315I-free survival probabilities. The model was validated with 820 patients from an additional 20 hospitals and demonstrated good discrimination, accuracy, and clinical net benefit. In addition, the nomogram effectively stratified patients into low-risk and high-risk categories, showing valuable clinical applicability.

This study revealed that the percentage of PBB was a predictor of the occurrence of the T315I mutation in CP-CML patients. An increased percentage of PBB indicates disease progression and is associated with higher grades of bone marrow fibrosis and difficulty in achieving treatment-free remission²¹. The incidence of ACA in CP-CML patients has been found to be approximately 5%²², similar to the findings of our study. ACAs are categorized into major and minor routes and have been shown to negatively affect survival and prognosis¹³. A German study revealed that ACA is common in the T315I mutation subtype, with the 3q26.2 rearrangement highly correlated with BCR::ABL1 mutations²³. Patients with genetically unstable CML are more likely to develop KD mutations, possibly related to changes in the DNA damage response, cell cycle checkpoints, and DNA repair^24,25. In CP-CML patients with the T315I mutation treated with ≥ 1 TKI, baseline data revealed that 73.3% of patients used dasatinib, whereas 4.4–57.8% of patients used other TKIs²⁶. The 2-year follow-up of the DASISION study revealed that among patients with BCR::ABL mutations in the dasatinib group, 70% had the T315I mutation, whereas no T315I mutation was found in the imatinib group²⁷. Our results are consistent with these findings, suggesting that a history of dasatinib use predicts a greater likelihood of the T315I mutation. Therefore, monitoring for the occurrence of the T315I mutation in patients currently using or who have previously used dasatinib is crucial. EMR has become an effective predictor of long-term progression-free survival (PFS) and OS⁴. Studies have shown that adjusting early intervention strategies on the basis of EMR at 3 months can be beneficial²⁸. A BCR::ABL1^IS ≤ 1% at 6 months is also a recognized predictor of long-term survival²⁹. In the IRIS study, the estimated 6-year PFS rate for patients with a BCR::ABL1^IS ≤ 1% at 6 months was 97%, whereas it was 80% for those with a BCR::ABL1^IS > 1% at 6 months. This finding supports the rationale that failure to achieve a molecular response is associated with the T315I mutation.

Additional mutations other than BCR::ABL1 KD mutations are gaining increasing attention. A position paper recommended the use of next-generation sequencing (NGS) for mutation detection in CP-CML patients experiencing "failure" in first- or second-line TKI treatment³⁰. Compared with Sanger sequencing, NGS can detect the T315I mutation three months earlier and identify mutations undetectable by Sanger sequencing³¹. In CML patients, the presence of epigenetic gene mutations at diagnosis, such as ASXL1, IKZF1, BCOR, TET1/2, IDH1/2, DNMT3A/3B, and EZH2, is associated with lower molecular response rates and reduced PFS^7,32,33. A novel HDAC I/IIb inhibitor has been shown to prevent T315I-induced CML progression and target leukemia stem cells (LSCs) responsible for CML relapse by inhibiting EZH2³⁴. ROS-induced oxidative DNA damage in LSCs can lead to the T315I mutation and IKZF1 deletion. Studies on the correlation between non-BCR::ABL1 KD mutations and T315I mutations are limited. We anticipate incorporating parameters related to gene mutations into our model in future research.

Our study has certain limitations. First, as a retrospective study, it inevitably has some degree of internal bias. Second, given the complexity of clinical situations, we did not strictly screen patients on the basis of the different first-line TKIs used or the various TKI switch scenarios, which may have reduced the model's accuracy but also increased its applicability. And it does not Impede the model’s ability to identify patients who require close monitoring.

In this study, we identified PBB, ACA, the choice of dasatinib as a TKI, a non-EMR at 3 months, and a BCR::ABL^IS > 1% at 6 months as predictors of the T315I mutation in CP-CML patients. Both the training set and the validation set demonstrated good performance of this nomogram. This visualized nomogram model provides clinicians with an intuitive tool to help identify patients requiring close monitoring. In the future, we will conduct prospective studies to further validate the accuracy of this model, and we will aim to incorporate clinical parameters during follow-up and gene data determined by NGS to optimize this nomogram model.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Acknowledgments

This study was performed by the CML Cooperation Group of Zhejiang Hematology and Zhejiang Clinical Medical Research Center of Hematology in China.

Funding

This research was funded by the Key R&D Program of Zhejiang (No. 2022C03137) and the Zhejiang Medical Association Clinical Medical Research Special Fund Project (No. 2022ZYC-D09).

Authorcontributions

H.-S.E., Y.-X.D., W.-D.J. and W.-Y. contributed equally to this work. H.-J., J.-J., and T.-H.Y. designed the study; H.-S.E., and Y.,-X.D. performed the analysis and wrote the draft; H.-S.E., W.-D.J., W.-Y., Z.-X.Q., F.-W.Y., Q.-H.L., L.-Y., C.-L.L., C.-L.H., L.-J., Z.-L., S.-Y.P., L.-L.L., T.-G.Y., Z.-H., C.-Y., Y.-X.F., F.-X.N., H.-L., Z.-K.F., X.-Y.M., X.-L.J., Z.-H.Q., Z.-G., J,-Y.Q., Z.-X.J., W.-G.Q., T.-J.M., Z.-F., Z.-Y.F., and Y.-G.L. collected the data; Y.,-X.D., W.-D.J., W.-Y. and Z.-J.J. aided in interpreting the results and verified the underlying study data; H.-J., J.-J., and T.-H.Y. supervised the work; and all the authors discussed the results and approved the final manuscript.

Competing interests

The authors declare no competing interests.

Faderl, S. et al. The biology of chronic myeloid leukemia. N Engl. J. Med. 341, 164–172. 10.1056/NEJM199907153410306 (1999).
Jabbour, E. & Kantarjian, H. Chronic myeloid leukemia: 2022 update on diagnosis, therapy, and monitoring. Am. J. Hematol. 97, 1236–1256. 10.1002/ajh.26642 (2022).
Brummendorf, T. H. et al. Bosutinib versus imatinib for newly diagnosed chronic phase chronic myeloid leukemia: final results from the BFORE trial. Leukemia. 36, 1825–1833. 10.1038/s41375-022-01589-y (2022).
Kantarjian, H. M. et al. Long-term outcomes with frontline nilotinib versus imatinib in newly diagnosed chronic myeloid leukemia in chronic phase: ENESTnd 10-year analysis. Leukemia. 35, 440–453. 10.1038/s41375-020-01111-2 (2021).
Senapati, J. et al. Management of chronic myeloid leukemia in 2023 - common ground and common sense. Blood Cancer J. 13, 58. 10.1038/s41408-023-00823-9 (2023).
Liu, J. et al. Recent advances in Bcr-Abl tyrosine kinase inhibitors for overriding T315I mutation. Chem. Biol. Drug Des. 97, 649–664. 10.1111/cbdd.13801 (2021).
Shah, N. P. et al. Chronic Myeloid Leukemia, Version 2.2024, NCCN Clinical Practice Guidelines in Oncology. J. Natl. Compr. Canc Netw. 22, 43–69. 10.6004/jnccn.2024.0007 (2024).
Al Hamad, M. Contribution of BCR-ABL molecular variants and leukemic stem cells in response and resistance to tyrosine kinase inhibitors: a review. F1000Res 10, 1288, doi: (2021). 10.12688/f1000research.74570.2
Haddad, F. G. et al. Characteristics and outcomes of patients with chronic myeloid leukemia and T315I mutation treated in the pre- and post-ponatinib era. Am. J. Hematol. 98, 1619–1626. 10.1002/ajh.27037 (2023).
Balachandran, V. P., Gonen, M., Smith, J. J. & DeMatteo, R. P. Nomograms in oncology: more than meets the eye. Lancet Oncol. 16, e173–180. 10.1016/S1470-2045(14)71116-7 (2015).
Hochhaus, A. et al. European LeukemiaNet 2020 recommendations for treating chronic myeloid leukemia. Leukemia. 34, 966–984. 10.1038/s41375-020-0776-2 (2020).
Fabarius, A. et al. Impact of additional cytogenetic aberrations at diagnosis on prognosis of CML: long-term observation of 1151 patients from the randomized CML Study IV. Blood. 118, 6760–6768. 10.1182/blood-2011-08-373902 (2011).
Hehlmann, R. et al. High-risk additional chromosomal abnormalities at low blast counts herald death by CML. Leukemia. 34, 2074–2086. 10.1038/s41375-020-0826-9 (2020).
Vickers, A. J., van Calster, B. & Steyerberg, E. W. A simple, step-by-step guide to interpreting decision curve analysis. Diagn. Progn Res. 3, 18. 10.1186/s41512-019-0064-7 (2019).
Nicolini, F. E. et al. Epidemiologic study on survival of chronic myeloid leukemia and Ph(+) acute lymphoblastic leukemia patients with BCR-ABL T315I mutation. Blood. 114, 5271–5278. 10.1182/blood-2009-04-219410 (2009).
Nicolini, F. E. et al. Overall survival with ponatinib versus allogeneic stem cell transplantation in Philadelphia chromosome-positive leukemias with the T315I mutation. Cancer. 123, 2875–2880. 10.1002/cncr.30558 (2017).
Soverini, S. et al. Prospective assessment of NGS-detectable mutations in CML patients with nonoptimal response: the NEXT-in-CML study. Blood. 135, 534–541. 10.1182/blood.2019002969 (2020).
Zhang, X. S. et al. Predictive scoring systems for molecular responses in persons with chronic phase chronic myeloid leukemia receiving initial imatinib therapy. Leukemia. 36, 2042–2049. 10.1038/s41375-022-01616-y (2022).
Krishnan, V. et al. A single-cell atlas identifies pretreatment features of primary imatinib resistance in chronic myeloid leukemia. Blood. 141, 2738–2755. 10.1182/blood.2022017295 (2023).
Zhang, X. S., Gale, R. P., Zhang, M. J., Huang, X. J. & Jiang, Q. A predictive scoring system for therapy-failure in persons with chronic myeloid leukemia receiving initial imatinib therapy. Leukemia. 36, 1336–1342. 10.1038/s41375-022-01527-y (2022).
Jacobi, H. et al. Myelofibrosis at diagnosis is associated with the failure of treatment-free remission in CML patients. Front. Pharmacol. 14, 1212392. 10.3389/fphar.2023.1212392 (2023).
Molica, M., Massaro, F. & Breccia, M. Diagnostic and prognostic cytogenetics of chronic myeloid leukaemia: an update. Expert Rev. Mol. Diagn. 17, 1001–1008. 10.1080/14737159.2017.1383156 (2017).
Chen, Z. et al. Cytogenetic landscape and impact in blast phase of chronic myeloid leukemia in the era of tyrosine kinase inhibitor therapy. Leukemia. 31, 585–592. 10.1038/leu.2016.231 (2017).
Muvarak, N., Nagaria, P. & Rassool, F. V. Genomic instability in chronic myeloid leukemia: targets for therapy? Curr. Hematol. Malig Rep. 7, 94–102. 10.1007/s11899-012-0119-0 (2012).
Popp, H. D. et al. DNA Damage and DNA Damage Response in Chronic Myeloid Leukemia. Int. J. Mol. Sci. 21 10.3390/ijms21041177 (2020).
Cortes, J. E. et al. Asciminib monotherapy in patients with chronic-phase chronic myeloid leukemia with the T315I mutation after >/=1 prior tyrosine kinase inhibitor: 2-year follow-up results. Leukemia. 10.1038/s41375-024-02278-8 (2024).
Kantarjian, H. M. et al. Dasatinib or imatinib in newly diagnosed chronic-phase chronic myeloid leukemia: 2-year follow-up from a randomized phase 3 trial (DASISION). Blood. 119, 1123–1129. 10.1182/blood-2011-08-376087 (2012).
Neelakantan, P. et al. Combining BCR-ABL1 transcript levels at 3 and 6 months in chronic myeloid leukemia: implications for early intervention strategies. Blood. 121, 2739–2742. 10.1182/blood-2012-11-466037 (2013).
Jabbour, E. et al. The achievement of an early complete cytogenetic response is a major determinant for outcome in patients with early chronic phase chronic myeloid leukemia treated with tyrosine kinase inhibitors. Blood. 118, 4541–4546. 10.1182/blood-2011-04-348110 (2011). quiz 4759.
Soverini, S. et al. Next-generation sequencing for BCR-ABL1 kinase domain mutation testing in patients with chronic myeloid leukemia: a position paper. J. Hematol. Oncol. 12, 131. 10.1186/s13045-019-0815-5 (2019).
Baer, C. et al. Ultra-deep sequencing leads to earlier and more sensitive detection of the tyrosine kinase inhibitor resistance mutation T315I in chronic myeloid leukemia. Haematologica. 101, 830–838. 10.3324/haematol.2016.145888 (2016).
Schonfeld, L. et al. ASXL1 mutations predict inferior molecular response to nilotinib treatment in chronic myeloid leukemia. Leukemia. 36, 2242–2249. 10.1038/s41375-022-01648-4 (2022).
Ochi, Y. et al. Clonal evolution and clinical implications of genetic abnormalities in blastic transformation of chronic myeloid leukaemia. Nat. Commun. 12, 2833. 10.1038/s41467-021-23097-w (2021).
Qiu, Q. et al. HDAC I/IIb selective inhibitor Purinostat Mesylate combined with GLS1 inhibition effectively eliminates CML stem cells. Bioact Mater. 21, 483–498. 10.1016/j.bioactmat.2022.08.006 (2023).

No competing interests reported.

FigS1.tif
Fig.S1. Kaplan-Meier curves in training and validation cohorts. Kaplan-Meier curves of T315I-free survival probability in chronic phase chronic myeloid leukemia patients in training and validation cohorts.
TableS1.docx

Download PDF

Editorial decision: Revision requested
22 Nov, 2024
Reviews received at journal
12 Nov, 2024
Reviewers agreed at journal
12 Nov, 2024
Reviews received at journal
01 Oct, 2024
Reviews received at journal
22 Sep, 2024
Reviewers agreed at journal
17 Sep, 2024
Reviewers agreed at journal
17 Sep, 2024
Reviewers invited by journal
17 Sep, 2024
Editor assigned by journal
06 Sep, 2024
Editor invited by journal
31 Aug, 2024
Submission checks completed at journal
29 Aug, 2024
First submitted to journal
20 Aug, 2024

You are reading this latest preprint version

A nomogram for predicting T315I-free survival in chronic phase chronic myeloid leukemia patients: a multicenter retrospective study

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Study design and participants

Ethics statement

Data collection and definition

Development and assessment of the nomogram

Statistical analysis

Results

Baseline characteristics

Nomogram model development

Nomogram model validation

Clinical application of the nomogram

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1