Surrogate-free machine learning-based organ dose reconstruction for pediatric abdominal radiotherapy.

To study radiotherapy-related adverse effects, detailed dose information (3D distribution) is needed for accurate dose-effect modeling. For childhood cancer survivors who underwent radiotherapy in the pre-CT era, only 2D radiographs were acquired, thus 3D dose distributions must be reconstructed from limited information. State-of-the-art methods achieve this by using 3D surrogate anatomies. These can however lack personalization and lead to coarse reconstructions. We present and validate a surrogate-free dose reconstruction method based on Machine Learning (ML). Abdominal planning CTs (n = 142) of recently-treated childhood cancer patients were gathered, their organs at risk were segmented, and 300 artificial Wilms' tumor plans were sampled automatically. Each artificial plan was automatically emulated on the 142 CTs, resulting in 42,600 3D dose distributions from which dose-volume metrics were derived. Anatomical features were extracted from digitally reconstructed radiographs simulated from the CTs to resemble historical radiographs. Further, patient and radiotherapy plan features typically available from historical treatment records were collected. An evolutionary ML algorithm was then used to link features to dose-volume metrics. Besides 5-fold cross validation, a further evaluation was done on an independent dataset of five CTs each associated with two clinical plans. Cross-validation resulted in mean absolute errors ≤ 0.6 Gy for organs completely inside or outside the field. For organs positioned at the edge of the field, mean absolute errors ≤ 1.7 Gy for [Formula: see text], ≤ 2.9 Gy for [Formula: see text], and ≤ 13% for [Formula: see text] and [Formula: see text], were obtained, without systematic bias. Similar results were found for the independent dataset. To conclude, we proposed a novel organ dose reconstruction method that uses ML models to predict dose-volume metric values given patient and plan features. Our approach is not only accurate, but also efficient, as the setup of a surrogate is no longer needed.

The heterogeneity of pediatric MLL-rearranged acute myeloid leukemia.

Translocations involving the mixed-lineage leukemia (MLL) gene, localized at 11q23, comprise 15 to 20% of all pediatric acute myeloid leukemia (AML) cases. This review summarizes current knowledge about the etiology, biology, clinical characteristics and differences in outcome in MLL-rearranged pediatric AML. Furthermore, we discuss the role of cooperating events in MLL-rearranged pediatric AML, and future therapeutic strategies to improve outcome. We conclude that MLL-rearranged pediatric AML is a heterogeneous disease, and prognosis depends on various factors, for example, translocation partner, age, WBC and additional cytogenetic aberrations. The relationship of outcome with specific translocation partners requires that they be searched for in the diagnostic work-up of AML. To achieve further improvements in outcome, unraveling the biology of MLL-rearranged pediatric AML is warranted.

High BRE expression in pediatric MLL-rearranged AML is associated with favorable outcome.

Translocations involving the mixed lineage leukemia (MLL) gene, localized at 11q23, frequently occur in pediatric acute myeloid leukemia (AML). We recently reported differences in prognosis between the different translocation partners, suggesting differences in biological background. To unravel the latter, we used microarrays to generate gene expression profiles of 245 pediatric AML cases, including 53 MLL-rearranged cases. Thereby, we identified a specific gene expression signature for t(9;11)(p22;q23), and identified BRE (brain and reproductive organ expressed) to be discriminative for t(9;11)(p22;q23) (P<0.001) when compared with other MLL subtypes. Patients with high BRE expression showed a significantly better 3-year relapse-free survival (pRFS) (80±13 vs 30±10%, P=0.02) within MLL-rearranged AML cases. Moreover, multivariate analysis identified high BRE expression as an independent favorable prognostic factor within pediatric AML for RFS (HR=0.2, P=0.04). No significant differences were identified for 3-year event-free survival or for 3-year overall survival. Forced expression of BRE did not result in altered cell proliferation, apoptosis or drug sensitivity, which could explain the favorable outcome. In conclusion, overexpression of the BRE gene is predominantly found in MLL-rearranged AML with t(9;11)(p22;q23). Although further investigation for the role of BRE in leukemogenesis and outcome is warranted, high BRE expression is an independent prognostic factor for pRFS in pediatric AML.

EVI1 overexpression in distinct subtypes of pediatric acute myeloid leukemia.

Overexpression of the ecotropic virus integration-1 (EVI1) gene (EVI1+), localized at chromosome 3q26, is associated with adverse outcome in adult acute myeloid leukemia (AML). In pediatric AML, 3q26 abnormalities are rare, and the role of EVI1 is unknown. We studied 228 pediatric AML samples for EVI1+ using gene expression profiling and RQ-PCR. EVI1+ was found in 20/213 (9%) of children with de novo AML, and in 4/8 with secondary AML. It was predominantly found in MLL-rearranged AML (13/47), monosomy 7 (2/3), or FAB M6/7 (6/10), and mutually exclusive with core-binding factor AML, t(15;17), and NPM1 mutations. Fluorescent in situ hybridization (FISH) was performed to detect cryptic 3q26 abnormalities. However, none of the EVI1+ patients harbored structural 3q26 alterations. Although significant differences in 4 years pEFS for EVI1+ and EVI1- pediatric AML were observed (28%+/-11 vs 44%+/-4, P=0.04), multivariate analysis did not identify EVI1+ as an independent prognostic factor. We conclude that EVI1+ can be found in approximately 10% of pediatric AML. Although EVI1+ was not an independent prognostic factor, it was predominantly found in subtypes of pediatric AML that are related with an intermediate to unfavorable prognosis. Further research should explain the role of EVI1+ in disease biology in these cases. Remarkably, no 3q26 abnormalities were identified in EVI1+ pediatric AML.

New insights to the MLL recombinome of acute leukemias.

Chromosomal rearrangements of the human MLL gene are associated with high-risk pediatric, adult and therapy-associated acute leukemias. These patients need to be identified, treated appropriately and minimal residual disease was monitored by quantitative PCR techniques. Genomic DNA was isolated from individual acute leukemia patients to identify and characterize chromosomal rearrangements involving the human MLL gene. A total of 760 MLL-rearranged biopsy samples obtained from 384 pediatric and 376 adult leukemia patients were characterized at the molecular level. The distribution of MLL breakpoints for clinical subtypes (acute lymphoblastic leukemia, acute myeloid leukemia, pediatric and adult) and fused translocation partner genes (TPGs) will be presented, including novel MLL fusion genes. Combined data of our study and recently published data revealed 104 different MLL rearrangements of which 64 TPGs are now characterized on the molecular level. Nine TPGs seem to be predominantly involved in genetic recombinations of MLL: AFF1/AF4, MLLT3/AF9, MLLT1/ENL, MLLT10/AF10, MLLT4/AF6, ELL, EPS15/AF1P, MLLT6/AF17 and SEPT6, respectively. Moreover, we describe for the first time the genetic network of reciprocal MLL gene fusions deriving from complex rearrangements.