Can Machine Learning Models Predict Asparaginase-associated Pancreatitis in Childhood Acute Lymphoblastic Leukemia

Asparaginase-associated pancreatitis (AAP) frequently affects children treated for acute lymphoblastic leukemia (ALL) causing severe acute and persisting complications. Known risk factors such as asparaginase dosing, older age and single nucleotide polymorphisms (SNPs) have insufficient odds ratios to allow personalized asparaginase therapy. In this study, we explored machine learning strategies for prediction of individual AAP risk. We integrated information on age, sex, and SNPs based on Illumina Omni2.5exome-8 arrays of patients with childhood ALL (N=1564, 244 with AAP 1.0 to 17.9 yo) from 10 international ALL consortia into machine learning models including regression, random forest, AdaBoost and artificial neural networks. A model with only age and sex had area under the receiver operating characteristic curve (ROC-AUC) of 0.62. Inclusion of 6 pancreatitis candidate gene SNPs or 4 validated pancreatitis SNPs boosted ROC-AUC somewhat (0.67) while 30 SNPs, identified through our AAP genome-wide association study cohort, boosted performance (0.80). Most predictive features included rs10273639 (PRSS1-PRSS2), rs10436957 (CTRC), rs13228878 (PRSS1/PRSS2), rs1505495 (GALNTL6), rs4655107 (EPHB2) and age (1 to 7 y). Second AAP following asparaginase re-exposure was predicted with ROC-AUC: 0.65. The machine learning models assist individual-level risk assessment of AAP for future prevention trials, and may legitimize asparaginase re-exposure when AAP risk is predicted to be low.

Impact of acute lymphoblastic leukemia induction therapy: findings from metabolomics on non-fasted plasma samples from a biorepository

INTRODUCTION

Acute lymphoblastic leukemia (ALL) is among the most common cancers in children. With improvements in combination chemotherapy regimens, the overall survival has increased to over 90%. However, the current challenge is to mitigate adverse events resulting from the complex therapy. Several chemotherapies intercept cancer metabolism, but little is known about their collective role in altering host metabolism.

OBJECTIVES

We profiled the metabolomic changes in plasma of ALL patients initial- and post- induction therapy.

METHODS

We exploited a biorepository of non-fasted plasma samples derived from the Dana Farber Cancer Institute ALL Consortium; these samples were obtained from 50 ALL patients initial- and post-induction therapy. Plasma metabolites and complex lipids were analyzed by high resolution tandem mass spectrometry and differential mobility tandem mass spectrometry. Data were analyzed using a covariate-adjusted regression model with multiplicity adjustment. Pathway enrichment analysis and co-expression network analysis were performed to identify unique clusters of molecules.

RESULTS

More than 1200 metabolites and complex lipids were identified in the total of global metabolomics and lipidomics platforms. Over 20% of those molecules were significantly altered. In the pathway enrichment analysis, lipids, particularly phosphatidylethanolamines (PEs), were identified. Network analysis indicated that the bioactive fatty acids, docosahexaenoic acid (DHA)-containing (22:6) triacylglycerols (TAGs), were decreased in the post-induction therapy.

CONCLUSION

Metabolomic profiling in ALL patients revealed a large number of alterations following induction chemotherapy. In particular, lipid metabolism was substantially altered. The changes in metabolites and complex lipids following induction therapy could provide insight into the adverse events experienced by ALL patients.

Validation of distinct type 2 diabetes clusters and their association with diabetes complications in the DEVOTE, LEADER and SUSTAIN-6 cardiovascular outcomes trials

AIMS

To validate the clusters of Swedish individuals with recent-onset diabetes at differential risk of complications, which were identified in a previous study, in three global populations with long-standing type 2 diabetes (T2D) who were at high cardiovascular risk, and to test for differences in the risk of major diabetes complications and survival endpoints.

MATERIALS AND METHODS

We assigned participants from recent global outcomes trials (DEVOTE [n = 7637], LEADER [n = 9340] and SUSTAIN-6 [n = 3297]) to the previously defined clusters according to age at diabetes diagnosis, baseline glycated haemoglobin (HbA1c) and body mass index (BMI). Outcomes were assessed using Kaplan-Meier analysis and log-rank tests.

RESULTS

The T2D clusters were consistently replicated across the three trial cohorts. The risk of major adverse cardiovascular events and cardiovascular death differed significantly, in all trials, across clusters over a median follow-up duration of 2.0, 3.8 and 2.1 years, respectively, and was highest for the cluster of participants with high HbA1c and low BMI (P < 0.05 in DEVOTE and LEADER). In LEADER and SUSTAIN-6, the risk of nephropathy differed across clusters (P < 0.0001 and P = 0.003, respectively). The risk of severe hypoglycaemia differed in DEVOTE (P = 0.006).

CONCLUSIONS

Previously identified clusters can be replicated in three geographically diverse cohorts of long-standing T2D and are associated with cluster-specific risk profiles for additional clinical and survival outcomes, providing further validation of the clustering methodology. The external validity and stability of clusters across cohorts provides a premise for future work to optimize the clustering approach to yield T2D subgroups with maximum predictive validity who may benefit from subtype-specific treatment paradigms.

Genome-Wide Association Meta-Analysis of Single-Nucleotide Polymorphisms and Symptomatic Venous Thromboembolism during Therapy for Acute Lymphoblastic Leukemia and Lymphoma in Caucasian Children

Symptomatic venous thromboembolism (VTE) occurs in five percent of children treated for acute lymphoblastic leukemia (ALL), but whether a genetic predisposition exists across different ALL treatment regimens has not been well studied.

METHODS

We undertook a genome-wide association study (GWAS) meta-analysis for VTE in consecutively treated children in the Nordic/Baltic acute lymphoblastic leukemia 2008 (ALL2008) cohort and the Australian Evaluation of Risk of ALL Treatment-Related Side-Effects (ERASE) cohort. A total of 92 cases and 1481 controls of European ancestry were included.

RESULTS

No SNPs reached genome-wide significance (p < 5 × 10-8) in either cohort. Among the top 34 single-nucleotide polymorphisms (SNPs) (p < 1 × 10-6), two loci had concordant effects in both cohorts: ALOX15B (rs1804772) (MAF: 1%; p = 3.95 × 10-7) that influences arachidonic acid metabolism and thus platelet aggregation, and KALRN (rs570684) (MAF: 1%; p = 4.34 × 10-7) that has been previously associated with risk of ischemic stroke, atherosclerosis, and early-onset coronary artery disease.

CONCLUSION

This represents the largest GWAS meta-analysis conducted to date associating SNPs to VTE in children and adolescents treated on childhood ALL protocols. Validation of these findings is needed and may then lead to patient stratification for VTE preventive interventions. As VTE hemostasis involves multiple pathways, a more powerful GWAS is needed to detect combination of variants associated with VTE.

Asparaginase-Associated Pancreatitis in Acute Lymphoblastic Leukemia: Results From the NOPHO ALL2008 Treatment of Patients 1-45 Years of Age

PURPOSE

Asparaginase-associated pancreatitis (AAP) is common in patients with acute lymphoblastic leukemia (ALL), but risk differences across age groups both in relation to first-time AAP and after asparaginase re-exposure have not been explored.

PATIENTS AND METHODS

We prospectively registered AAP (n = 168) during treatment of 2,448 consecutive ALL patients aged 1.0-45.9 years diagnosed from July 2008 to October 2018 and treated according to the Nordic Society of Pediatric Hematology and Oncology (NOPHO) ALL2008 protocol.

RESULTS

Compared with patients aged 1.0-9.9 years, adjusted AAP hazard ratios (HRa) were associated with higher age with almost identical HRa (1.6; 95% CI, 1.1 to 2.3; P = .02) for adolescents (10.0-17.9 years) and adults (18.0-45.9 years). The day 280 cumulative incidences of AAP were 7.0% for children (1.0-9.9 years: 95% CI, 5.4 to 8.6), 10.1% for adolescents (10.0 to 17.9 years: 95% CI, 7.0 to 13.3), and 11.0% for adults (18.0-45.9 years: 95% CI, 7.1 to 14.9; P = .03). Adolescents had increased odds of both acute (odds ratio [OR], 5.2; 95% CI, 2.1 to 13.2; P = .0005) and persisting complications (OR, 6.7; 95% CI, 2.4 to 18.4; P = .0002) compared with children (1.0-9.9 years), whereas adults had increased odds of only persisting complications (OR, 4.1; 95% CI, 1.4 to 11.8; P = .01). Fifteen of 34 asparaginase-rechallenged patients developed a second AAP. Asparaginase was truncated in 17/21 patients with AAP who subsequently developed leukemic relapse, but neither AAP nor the asparaginase truncation was associated with increased risk of relapse.

CONCLUSION

Older children and adults had similar AAP risk, whereas morbidity was most pronounced among adolescents. Asparaginase re-exposure should be considered only for patients with an anticipated high risk of leukemic relapse, because multiple studies strongly indicate that reduction of asparaginase treatment intensity increases the risk of relapse.

Intermittent Versus Continuous PEG-Asparaginase to Reduce Asparaginase-Associated Toxicities: A NOPHO ALL2008 Randomized Study

PURPOSE

Asparaginase is an essential drug in childhood acute lymphoblastic leukemia (ALL) therapy and is frequently given for months to obtain continuous asparagine depletion. We randomly assigned patients to continuous versus intermittent pegylated-asparaginase (PEG-asp) treatment, hypothesizing there would be decreased toxicity with unchanged efficacy.

METHODS

Children (median age, 4.2 years) treated for non–high-risk ALL according to the Nordic Society for Pediatric Hematology and Oncology ALL2008 protocol received five intramuscular PEG-asp injections (1,000 IU/m2) every two weeks and were then randomly assigned to additional three doses (6-week intervals [experimental arm], n = 309) versus 10 doses (2-week intervals [standard arm], n = 316). The primary end point was noninferior (6% margin) disease-free survival. Toxicity reduction was a secondary end point. Occurrence of asparaginase-associated hypersensitivity, pancreatitis, osteonecrosis, and thromboembolism were prospectively registered.

RESULTS

After a median follow-up of 4.1 years, the 5-year disease-free survival was 92.2% (95% CI, 88.6 to 95.8) and 90.8% (95% CI, 87.0 to 94.6) in the experimental and standard arms, respectively. The 3-year cumulative incidence of any first asparaginase-associated toxicity (hypersensitivity [n = 13]; osteonecrosis [n = 29]; pancreatitis [n = 24]; thromboembolism [n = 17]) was 9.3% in the experimental arm and 18.1% in the standard arm ( P = .001). Asparaginase-associated toxicity reduction was confirmed in sex- and risk-group–adjusted Cox regression analysis stratified by age (≥ 10 and < 10 years; hazard ratio, 0.48; P = .001). The experimental arm had the lowest incidences of all four toxicities, reaching significance for pancreatitis (6-month risk, 5.8% v 1.3%; P = .002).

CONCLUSION

The excellent cure rates and reduced toxicity risk support the use of intermittent PEG-asp therapy after the first 10 weeks in future childhood ALL trials that apply prolonged PEG-asp therapy.

Trypsin-encoding PRSS1-PRSS2 variations influence the risk of asparaginase-associated pancreatitis in children with acute lymphoblastic leukemia: a Ponte di Legno toxicity working group report

Asparaginase-associated pancreatitis is a life-threatening toxicity to childhood acute lymphoblastic leukemia treatment. To elucidate genetic predisposition and asparaginase-associated pancreatitis pathogenesis, ten trial groups contributed remission samples from patients aged 1.0-17.9 years treated for acute lymphoblastic leukemia between 2000 and 2016. Cases (n=244) were defined by the presence of at least two of the following criteria: (i) abdominal pain; (ii) levels of pancreatic enzymes ≥3 × upper normal limit; and (iii) imaging compatible with pancreatitis. Controls (n=1320) completed intended asparaginase therapy, with 78% receiving ≥8 injections of pegylated-asparaginase, without developing asparaginase-associated pancreatitis. rs62228256 on 20q13.2 showed the strongest association with the development of asparaginase-associated pancreatitis (odds ratio=3.75; P=5.2×10-8). Moreover, rs13228878 (OR=0.61; P=7.1×10-6) and rs10273639 (OR=0.62; P=1.1×10-5) on 7q34 showed significant association with the risk of asparaginase-associated pancreatitis. A Dana Farber Cancer Institute ALL Consortium cohort consisting of patients treated on protocols between 1987 and 2004 (controls=285, cases=33), and the Children's Oncology Group AALL0232 cohort (controls=2653, cases=76) were available as replication cohorts for the 20q13.2 and 7q34 variants, respectively. While rs62228256 was not validated as a risk factor (P=0.77), both rs13228878 (P=0.03) and rs10273639 (P=0.04) were. rs13228878 and rs10273639 are in high linkage disequilibrium (r2=0.94) and associated with elevated expression of the PRSS1 gene, which encodes for trypsinogen, and are known risk variants for alcohol-associated and sporadic pancreatitis in adults. Intra-pancreatic trypsinogen cleavage to proteolytic trypsin induces autodigestion and pancreatitis. In conclusion, this study finds a shared genetic predisposition between asparaginase-associated pancreatitis and non-asparaginase-associated pancreatitis, and targeting the trypsinogen activation pathway may enable identification of effective interventions for asparaginase-associated pancreatitis.

Genetic predisposition to PEG-asparaginase hypersensitivity in children treated according to NOPHO ALL2008

Asparaginase is essential in childhood acute lymphoblastic leukaemia (ALL) treatment, however hypersensitivity reactions to pegylated asparaginase (PEG-asparaginase) hampers anti-neoplastic efficacy. Patients with PEG-asparaginase hypersensitivity have been shown to possess zero asparaginase enzyme activity. Using this measurement to define the phenotype, we investigated genetic predisposition to PEG-asparaginase hypersensitivity in a genome-wide association study (GWAS). From July 2008 to March 2016, 1494 children were treated on the Nordic Society of Paediatric Haematology and Oncology ALL2008 protocol. Cases were defined by clinical hypersensitivity and no enzyme activity, controls had enzyme activity ≥ 100 iu/l and no hypersensitivity symptoms. PEG-asparaginase hypersensitivity was reported in 13·8% (206/1494) of patients. Fifty-nine cases and 772 controls fulfilled GWAS inclusion criteria. The CNOT3 variant rs73062673 on 19q13.42, was associated with PEG-asparaginase allergy (P = 4·68 × 10^-8 ). We further identified two signals on chromosome 6 in relation to HLA-DQA1 (P = 9·37 × 10^-6 ) and TAP2 (P = 1·59 × 10^-5 ). This study associated variants in CNOT3 and in the human leucocyte antigen (HLA) region with PEG-asparaginase hypersensitivity, suggesting that not only genetic variations in the HLA region, but also regulation of these genes are of importance in the biology of this toxicity. Furthermore, our study emphasizes the importance of using asparaginase enzyme activity measurements to identify PEG-asparaginase hypersensitivity.

Asparaginase-associated pancreatitis in childhood acute lymphoblastic leukaemia: an observational Ponte di Legno Toxicity Working Group study

BACKGROUND

Survival for childhood acute lymphoblastic leukaemia surpasses 90% with contemporary therapy; however, patients remain burdened by the severe toxic effects of treatment, including asparaginase-associated pancreatitis. To investigate the risk of complications and risk of re-exposing patients with asparaginase-associated pancreatitis to asparaginase, 18 acute lymphoblastic leukaemia trial groups merged data for this observational study.

METHODS

Patient files from 26 trials run by 18 trial groups were reviewed on children (aged 1·0-17·9 years) diagnosed with t(9;22)-negative acute lymphoblastic leukaemia between June 1, 1996, and Jan 1, 2016, who within 50 days of asparaginase exposure developed asparaginase-associated pancreatitis. Asparaginase-associated pancreatitis was defined by at least two criteria: abdominal pain, pancreatic enzymes at least three times the upper limit of normal (ULN), and imaging compatible with pancreatitis. Patients without sufficient data for diagnostic criteria were excluded. Primary outcomes were defined as acute and persisting complications of asparaginase-associated pancreatitis and risk of re-exposing patients who suffered an episode of asparaginase-associated pancreatitis to asparaginase. Data were collected from Feb 2, 2015, to June 30, 2016, and analysed and stored in a common database at Rigshospitalet, Copenhagen, Denmark.

FINDINGS

Of 465 patients with asparaginase-associated pancreatitis, 33 (8%) of 424 with available data needed mechanical ventilation, 109 (26%) of 422 developed pseudocysts, acute insulin therapy was needed in 81 (21%) of 393, and seven (2%) of 458 patients died. Risk of assisted mechanical ventilation, need for insulin, pseudocysts, or death was associated with older age (median age for patients with complications 10·5 years [IQR 6·4-13·8] vs without complications 6·1 years [IQR 3·6-12·2], p<0·0001), and having one or more affected vital signs (fever, hypotension, tachycardia, or tachypnoea; 96 [44%] of 217 patients with affected vital signs vs 11 [24%] of 46 patients without affected vital signs, p=0·02). 1 year after diagnosis of asparaginase-associated pancreatitis, 31 (11%) of 275 patients still needed insulin or had recurrent abdominal pain or both. Both the risk of persisting need for insulin therapy and recurrent abdominal pain were associated with having had pseudocysts (odds ratio [OR] 9·48 [95% CI 3·01-35·49], p=0·0002 for insulin therapy; OR 11·79 [4·30-37·98], p<0·0001 for recurrent abdominal pain). Within 8 years of asparaginase-associated pancreatitis, risk of abdominal symptoms dropped from 8% (26 of 312) to 0% (0 of 35) but the need for insulin therapy remained constant (9%, three of 35). 96 patients were re-exposed to asparaginase, including 59 after a severe asparaginase-associated pancreatitis (abdominal pain or pancreatic enzymes at least three times the ULN or both lasting longer than 72 h). 44 (46%) patients developed a second asparaginase-associated pancreatitis, 22 (52%) of 43 being severe. Risk of persisting need for insulin or abdominal pain after having had two versus one asparaginase-associated pancreatitis did not differ (three [7%] of 42 vs 28 [12%] of 233, p=0·51). Risk of a second asparaginase-associated pancreatitis was not associated with any baseline patient characteristics.

INTERPRETATION

Since the risk of a second asparaginase-associated pancreatitis was not associated with severity of the first asparaginase-associated pancreatitis and a second asparaginase-associated pancreatitis did not involve an increased risk of complications, asparaginase re-exposure should be determined mainly by the anticipated need for asparaginase for antileukaemic efficacy. A study of the genetic risk factors identifying patients in whom asparaginase exposure should be restricted is needed.

FUNDING

The Danish Childhood Cancer Foundation and The Danish Cancer Society (R150-A10181).

Asparaginase-associated pancreatitis: a study on phenotype and genotype in the NOPHO ALL2008 protocol

Asparaginase (ASP)-associated pancreatitis (AAP) occurs during acute lymphoblastic leukemia treatment. Among 1285 children (1.0-17.9 years) diagnosed during July 2008-December 2014 and treated according to the Nordic/Baltic ALL2008 protocol, 86 (cumulative incidence=6.8%) developed AAP. Seventy-three cases were severe (diagnostic AAP criteria persisting >72 h) and 13 mild. Cases were older than controls (median: 6.5 vs 4.5 years; P=0.001). Pseudocysts developed in 28%. Of the 20 re-exposed to ASP, 9 (45%) developed a second AAP. After a median follow-up of 2.3 years, 8% needed permanent insulin therapy, and 7% had recurrent abdominal pain. Germline DNA on 62 cases and 638 controls was genotyped on Omni2.5exome-8-v1.2 BeadChip arrays. Overall, the ULK2 variant rs281366 showed the strongest association with AAP (P=5.8 × 10^-7; odds ratio (OR)=6.7). Cases with the rs281366 variant were younger (4.3 vs 8 years; P=0.015) and had lower risk of AAP-related complications (15% vs 43%; P=0.13) compared with cases without this variant. Among 45 cases and 517 controls <10 years, the strongest associations with AAP were found for RGS6 variant rs17179470 (P=9.8 × 10^-9; OR=7.3). Rs281366 is located in the ULK2 gene involved in autophagy, and RGS6 regulates G-protein signaling regulating cell dynamics. More than 50% of AAP cases <10 years carried one or both risk alleles.