A Clinical Prognostic Model Based on Machine Learning from the Fondazione Italiana Linfomi (FIL) MCL0208 Phase III Trial.

BACKGROUND: Multicenter clinical trials are producing growing amounts of clinical data. Machine Learning (ML) might facilitate the discovery of novel tools for prognostication and disease-stratification. Taking advantage of a systematic collection of multiple variables, we developed a model derived from data collected on 300 patients with mantle cell lymphoma (MCL) from the Fondazione Italiana Linfomi-MCL0208 phase III trial (NCT02354313). METHODS: We developed a score with a clustering algorithm applied to clinical variables. The candidate score was correlated to overall survival (OS) and validated in two independent data series from the European MCL Network (NCT00209222, NCT00209209); Results: Three groups of patients were significantly discriminated: Low, Intermediate (Int), and High risk (High). Seven discriminants were identified by a feature reduction approach: albumin, Ki-67, lactate dehydrogenase, lymphocytes, platelets, bone marrow infiltration, and B-symptoms. Accordingly, patients in the Int and High groups had shorter OS rates than those in the Low and Int groups, respectively (Int→Low, HR: 3.1, 95% CI: 1.0-9.6; High→Int, HR: 2.3, 95% CI: 1.5-4.7). Based on the 7 markers, we defined the engineered MCL international prognostic index (eMIPI), which was validated and confirmed in two independent cohorts; Conclusions: We developed and validated a ML-based prognostic model for MCL. Even when currently limited to baseline predictors, our approach has high scalability potential.

Impact of different exposure models and spatial resolution on the long-term effects of air pollution.

Long-term exposure to air pollution has been related to mortality in several epidemiological studies. The investigations have assessed exposure using various methods achieving different accuracy in predicting air pollutants concentrations. The comparison of the health effects estimates are therefore challenging. This paper aims to compare the effect estimates of the long-term effects of air pollutants (particulate matter with aerodynamic diameter less than 10 µm, PM10, and nitrogen dioxide, NO2) on cause-specific mortality in the Rome Longitudinal Study, using exposure estimates obtained with different models and spatial resolutions. Annual averages of NO2 and PM10 were estimated for the year 2015 in a large portion of the Rome urban area (12 × 12 km2) applying three modelling techniques available at increasing spatial resolution: 1) a chemical transport model (CTM) at 1km resolution; 2) a land-use random forest (LURF) approach at 200m resolution; 3) a micro-scale Lagrangian particle dispersion model (PMSS) taking into account the effect of buildings structure at 4 m resolution with results post processed at different buffer sizes (12, 24, 52, 100 and 200 m). All the exposures were assigned at the residential addresses of 482,259 citizens of Rome 30+ years of age who were enrolled on 2001 and followed-up till 2015. The association between annual exposures and natural-cause, cardiovascular (CVD) and respiratory (RESP) mortality were estimated using Cox proportional hazards models adjusted for individual and area-level confounders. We found different distributions of both NO2 and PM10 concentrations, across models and spatial resolutions. Natural cause and CVD mortality outcomes were all positively associated with NO2 and PM10 regardless of the model and spatial resolution when using a relative scale of the exposure such as the interquartile range (IQR): adjusted Hazard Ratios (HR), and 95% confidence intervals (CI), of natural cause mortality, per IQR increments in the two pollutants, ranged between 1.012 (1.004, 1.021) and 1.018 (1.007, 1.028) for the different NO2 estimates, and between 1.010 (1.000, 1.020) and 1.020 (1.008, 1.031) for PM10, with a tendency of larger effect for lower resolution exposures. The latter was even stronger when a fixed value of 10 µg/m3 is used to calculate HRs. Long-term effects of air pollution on mortality in Rome were consistent across different models for exposure assessment, and different spatial resolutions.

Immunoglobulin kappa deleting element rearrangements are candidate targets for minimal residual disease evaluation in mantle cell lymphoma.

Minimal residual disease (MRD) assessment is of high clinical relevance in patients with mantle cell lymphoma (MCL). In mature B-cell malignancies, the presence of somatic hypermutations (SHM) in Variable-Diversity-Joining Heavy chain (VDJH) rearrangements leads to frequent mismatches between primers, probes, and the target, thus impairing tumor cells quantification. Alternative targets, such as immunoglobulin kappa-deleting-element (IGK-Kde) rearrangements, might be suitable for MRD detection. We aimed at evaluating the applicability of IGK-Kde rearrangements for MRD quantification in MCL patients by real-time quantitative polymerase chain reaction (RQ-PCR)/digital-droplet-PCR (ddPCR). IGK screening was performed on bone marrow samples from two cohorts: the first from Turin (22 patients enrolled in the FIL-MCL0208 trial, NCT02354313) and the second from Rome (15 patients). IGK-Kde rearrangements were found in 76% (28/37) of cases, representing the sole molecular marker in 73% (8/11) of IGH-BCL1/IGH negative cases. MRD RQ-PCR monitoring was possible in 57% (16/28) of cases, showing a 100% concordance with the conventional targets. However, the frequent background amplification affected the sensitivity of the assay, that was lower in MCL compared to acute lymphoblastic leukemia and in line with multiple myeloma published results. ddPCR had a good concordance with RQ-PCR and it might help to identify false positive/negative results. From a clinical perspective, we suggest that IGK-Kde can be a candidate target for MRD monitoring and deserves a validation of its predictive value in prospective MCL series.

KMT2D mutations and TP53 disruptions are poor prognostic biomarkers in mantle cell lymphoma receiving high-dose therapy: a FIL study.

In recent years, the outcome of mantle cell lymphoma (MCL) has improved, especially in younger patients, receiving cytarabine-containing chemoimmunotherapy and autologous stem cell transplantation. Nevertheless, a proportion of MCL patients still experience early failure. To identify biomarkers anticipating failure of intensive chemotherapy in MCL, we performed target resequencing and DNA profiling of purified tumor samples collected from patients enrolled in the prospective FIL-MCL0208 phase 3 trial (high-dose chemoimmunotherapy followed by autologous transplantation and randomized lenalidomide maintenance). Mutations of KMT2D and disruption of TP53 by deletion or mutation associated with an increased risk of progression and death, both in univariate and multivariate analysis. By adding KMT2D mutations and TP53 disruption to the MIPI-c backbone, we derived a new prognostic index, the "MIPI-genetic" ("MIPI- g"). The "MIPI-g" improved the model discrimination ability compared to the MIPI-c alone, defining three risk groups: i) low-risk patients (4-year progression free survival and overall survival of 72.0% and 94.5%); ii) inter-mediate-risk patients (4-year progression free survival and overall survival of 42.2% and 65.8%) and iii) high-risk patients (4-year progression free survival and overall survival of 11.5% and 44.9%). Our results: i) confirm that TP53 disruption identifies a high-risk population characterized by poor sensitivity to conventional or intensified chemotherapy; ii) provide the pivotal evidence that patients harboring KMT2D mutations share the same poor outcome as patients harboring TP53 disruption; and iii) allow to develop a tool for the identification of high-risk MCL patients for whom novel therapeutic strategies need to be investigated. (Trial registered at clinicaltrials.gov identifier: NCT02354313).

Applying Data Warehousing to a Phase III Clinical Trial From the Fondazione Italiana Linfomi Ensures Superior Data Quality and Improved Assessment of Clinical Outcomes.

PURPOSE: Data collection in clinical trials is becoming complex, with a huge number of variables that need to be recorded, verified, and analyzed to effectively measure clinical outcomes. In this study, we used data warehouse (DW) concepts to achieve this goal. A DW was developed to accommodate data from a large clinical trial, including all the characteristics collected. We present the results related to baseline variables with the following objectives: developing a data quality (DQ) control strategy and improving outcome analysis according to the clinical trial primary end points. METHODS: Data were retrieved from the electronic case reporting forms (eCRFs) of the phase III, multicenter MCL0208 trial (ClinicalTrials.gov identifier: NCT02354313) of the Fondazione Italiana Linfomi for younger patients with untreated mantle cell lymphoma (MCL). The DW was created with a relational database management system. Recommended DQ dimensions were observed to monitor the activity of each site to handle DQ management during patient follow-up. The DQ management was applied to clinically relevant parameters that predicted progression-free survival to assess its impact. RESULTS: The DW encompassed 16 tables, which included 226 variables for 300 patients and 199,500 items of data. The tool allowed cross-comparison analysis and detected some incongruities in eCRFs, prompting queries to clinical centers. This had an impact on clinical end points, as the DQ control strategy was able to improve the prognostic stratification according to single parameters, such as tumor infiltration by flow cytometry, and even using established prognosticators, such as the MCL International Prognostic Index. CONCLUSION: The DW is a powerful tool to organize results from large phase III clinical trials and to effectively improve DQ through the application of effective engineered tools.

Minimal residual disease (MRD) in non-Hodgkin lymphomas: Interlaboratory reproducibility on marrow samples with very low levels of disease within the FIL (Fondazione Italiana Linfomi) MRD Network.

In 2009, the four laboratories of the Fondazione Italiana Linfomi (FIL) minimal residual disease (MRD) Network started a collaborative effort to harmonize and standardize their methodologies at the national level, performing quality control (QC) rounds for follicular lymphoma (FL) and mantle cell lymphoma (MCL) MRD assessment. In 16 QC rounds between 2010 and 2017, the four laboratories received 208 bone marrow (BM) samples (126 FL; 82 MCL); 187 were analyzed, according to the EuroMRD Consortium guidelines, by both nested (NEST) polymerase chain reaction (PCR) and real-time quantitative (RQ) PCR for BCL2/IGH MBR or IGHV rearrangements. Here, we aimed at analyzing the samples that challenged the interlaboratory reproducibility and data interpretation. Overall, 156/187 BM samples (83%) were concordantly classified as NEST+/RQ+ or NEST-/RQ- by all the four laboratories. The remaining 31 samples (17%) resulted alternatively positive and negative in the interlaboratory evaluations, independently of the method and the type of rearrangement, and were defined "borderline" (brd) samples: 12 proved NEST brd/RQ brd, 7 NEST-/RQ brd, 10 NEST brd/RQ positive not quantifiable (PNQ), and 2 NEST brd/RQ-. Results did not change even increasing the number of replicates/sample. In 6/31 brd samples, droplet digital PCR (ddPCR) was tested and showed no interlaboratory discordance. Despite the high interlaboratory reproducibility in the MRD analysis obtained and maintained by the QC round strategy, samples with the lowest MRD levels can still represent a challenge: 17% (31/187) of our samples showed discordant results in interlaboratory assessments, with 6.4% (12/187) remained brd even applying the two methods. Thus, although representing a minority, brd samples are still problematic, especially when a clinically oriented interpretation of MRD results is required. Alternative, novel methods such as ddPCR and next-generation sequencing have the potential to overcome the current limitations.

HashClone: a new tool to quantify the minimal residual disease in B-cell lymphoma from deep sequencing data.

BACKGROUND: Mantle Cell Lymphoma (MCL) is a B cell aggressive neoplasia accounting for about the 6% of all lymphomas. The most common molecular marker of clonality in MCL, as in other B lymphoproliferative disorders, is the ImmunoGlobulin Heavy chain (IGH) rearrangement, occurring in B-lymphocytes. The patient-specific IGH rearrangement is extensively used to monitor the Minimal Residual Disease (MRD) after treatment through the standardized Allele-Specific Oligonucleotides Quantitative Polymerase Chain Reaction based technique. Recently, several studies have suggested that the IGH monitoring through deep sequencing techniques can produce not only comparable results to Polymerase Chain Reaction-based methods, but also might overcome the classical technique in terms of feasibility and sensitivity. However, no standard bioinformatics tool is available at the moment for data analysis in this context. RESULTS: In this paper we present HashClone, an easy-to-use and reliable bioinformatics tool that provides B-cells clonality assessment and MRD monitoring over time analyzing data from Next-Generation Sequencing (NGS) technique. The HashClone strategy-based is composed of three steps: the first and second steps implement an alignment-free prediction method that identifies a set of putative clones belonging to the repertoire of the patient under study. In the third step the IGH variable region, diversity region, and joining region identification is obtained by the alignment of rearrangements with respect to the international ImMunoGenetics information system database. Moreover, a provided graphical user interface for HashClone execution and clonality visualization over time facilitate the tool use and the results interpretation. The HashClone performance was tested on the NGS data derived from MCL patients to assess the major B-cell clone in the diagnostic samples and to monitor the MRD in the real and artificial follow up samples. CONCLUSIONS: Our experiments show that in all the experimental settings, HashClone was able to correctly detect the major B-cell clones and to precisely follow them in several samples showing better accuracy than the state-of-art tool.

Rituximab, bendamustine, and low-dose cytarabine as induction therapy in elderly patients with mantle cell lymphoma: a multicentre, phase 2 trial from Fondazione Italiana Linfomi.

BACKGROUND: The combination of rituximab, bendamustine, and cytarabine (R-BAC) was highly active in a pilot trial of mantle cell lymphoma, but its use was restricted by high haematological toxicity. We aimed to assess the efficacy and safety of an R-BAC regimen with low-dose cytarabine (RBAC500). METHODS: In this multicentre, phase 2 trial, we recruited previously untreated patients with an established histological diagnosis of mantle cell lymphoma from 29 Fondazione Italiana Linfomi centres in Italy. Patients had to be older than 65 years and fit according to the comprehensive geriatric assessment, or aged 60-65 years if they were ineligible for high-dose chemotherapy plus autologous stem-cell transplantation and were fit or unfit. All patients received RBAC500 (rituximab 375 mg/m2 on day 1, bendamustine 70 mg/m2 on days 2 and 3, and cytarabine 500 mg/m2 on days 2-4; all administered intravenously) every 4 weeks for up to six cycles. Primary endpoints were the proportion of patients achieving complete response at the end of treatment and toxicity, defined as the occurrence of any of the stop treatment criteria or of any episode of relevant toxicity. All patients who started at least one cycle of RBAC500 were included in the primary and safety analyses. Using efficacy and toxicity as a composite primary endpoint, we considered the final conclusion positive if more than 28 of 57 patients achieve complete response and fewer than 18 of 57 patients report toxicities. This study is registered with EudraCT, number 2011-005739-23, and ClinicalTrials.gov, number NCT01662050, and is completed. FINDINGS: Between May 2, 2012, and Feb 25, 2014, we enrolled 57 patients (median age 71 years, IQR 67-75). 54 (95%) patients received at least four RBAC500 cycles (three discontinued because of toxicity), and 38 (67%) completed six cycles. Two (4%) had disease progression (one after the fourth cycle and one after the sixth cycle). All 52 (91%, lower limit of one-sided 95% CI 85%) remaining patients achieved complete response at the end of treatment. 23 (40%, upper limit of one-sided 95% CI 53%) of 57 patients had at least one episode of relevant toxicity. The most frequent grade 3-4 haematological toxicities were neutropenia (149 [49%] of 304 cycles) and thrombocytopenia (158 [52%]). Most treatment-related non-haematological adverse events were of grade 1-2, with the most frequent ones being fatigue (14 [25%] patients), nausea or vomiting (12 [21%]), and infusion-related reactions or tumour lysis syndrome (12 [21%]). 41 (72%) patients required a dose reduction. 12 patients died during the study, but no deaths were related to treatment. INTERPRETATION: RBAC500 is an effective treatment for elderly patients with mantle cell lymphoma and, despite not meeting our prespecified safety boundary, haematological toxicity was manageable with appropriate supportive care and dose reduction. Since maintenance therapy is not required, RBAC500 could be considered an option and should be studied in phase 3 trials. FUNDING: Fondazione Italiana Linfomi and Mundipharma.

Minimal Residual Disease Detection by Droplet Digital PCR in Multiple Myeloma, Mantle Cell Lymphoma, and Follicular Lymphoma: A Comparison with Real-Time PCR.

Real-time quantitative PCR (qPCR) is a well-established tool for minimal residual disease (MRD) detection in mature lymphoid malignancies. Despite remarkable sensitivity and specificity, qPCR has some limitations, particularly in the need for a reference standard curve, based on target serial dilutions. In this study, we established droplet digital PCR (ddPCR) for MRD monitoring in multiple myeloma, mantle cell lymphoma, and follicular lymphoma and compared it head-to-head with qPCR. We observed that ddPCR has sensitivity, accuracy, and reproducibility comparable with qPCR. We then compared the two approaches in 69 patients with a documented molecular marker at diagnosis (18 multiple myelomas, 21 mantle cell lymphomas assessed with the immunoglobulin gene rearrangement, and 30 follicular lymphomas with the use of the BCL2/immunoglobulin gene major breakpoint region rearrangement). ddPCR was successful in 100% of cases, whereas qPCR failed to provide a reliable standard curve in three patients. Overall, 222 of 225 samples were evaluable by both methods. The comparison highlighted a good concordance (r = 0.94, P < 0.0001) with 189 of 222 samples (85.1%; 95% CI, 80.4%-89.8%) being fully concordant. We found that ddPCR is a reliable tool for MRD detection with greater applicability and reduced labor intensiveness than qPCR. It will be necessary to authorize ddPCR as an outcome predictor tool in controlled clinical settings and multilaboratory standardization programs.

Comparison of two real-time quantitative polymerase chain reaction strategies for minimal residual disease evaluation in lymphoproliferative disorders: correlation between immunoglobulin gene mutation load and real-time quantitative polymerase chain reaction performance.

We compared two strategies for minimal residual disease evaluation of B-cell lymphoproliferative disorders characterized by a variable immunoglobulin heavy chain (IGH) genes mutation load. Twenty-five samples from chronic lymphocytic leukaemia (n = 18) or mantle cell lymphoma (n = 7) patients were analyzed. Based on IGH variable region genes, 22/25 samples carried > 2% mutations, 20/25 > 5%. In the IGH joining region genes, 23/25 samples carried > 2% mutations, 18/25 > 5%. Real-time quantitative polymerase chain reaction was performed on IGH genes using two strategies: method A utilizes two patient-specific primers, whereas method B employs one patient-specific and one germline primer, with different positions on the variable, diversity and joining regions. Twenty-three samples (92%) resulted evaluable using method A, only six (24%) by method B. Method B poor performance was specifically evident among mutated IGH variable/joining region cases, although no specific mutation load above, which the real-time quantitative polymerase chain reaction failed was found. The molecular strategies for minimal residual disease evaluation should be adapted to the B-cell receptor features of the disease investigated.

Rituximab-based pre-emptive treatment of molecular relapse in follicular and mantle cell lymphoma.

Pre-emptive rituximab (pRTX) might represent an effective approach for patients with follicular (FL) and mantle cell lymphoma (MCL) experiencing molecular relapse (M-rel). However, available experience is still limited. We retrospectively collected FL and MCL cases that underwent pRTX with four weekly rituximab infusions (375 mg/m²) due to molecular persistence or M-rel. M-rel was assessed using nested polymerase chain reaction (PCR) and real-time quantitative PCR using the Bcl-1/IGH, Bcl-2/IGH or the immunoglobulin heavy chain rearrangement. Twenty-three occurrences of M-rel or persistence were treated in 18 patients (nine MCL and nine FL). The pRTX reinduced molecular remission (MR) in 17/23 cases (7/9 FL and 10/14 MCL). The median time to MR reinduction was 4.5 months (range 3-12), and the median duration of the first MR reinduction was 34 months (range 12-72). In five MCL cases, pRTX was used to treat subsequent M-rels, with success in four cases. No clinical relapses were seen within 2 years of successful reinduction of MR. Progression-free survival after pRTX was 64 % at a median follow-up of 6 years. pRTX was feasible and safe and effectively reinduced MR in FL and MCL patients (74 %). Prospective trials are needed to verify the clinical benefit of similar approaches.

Clinical implications and prognostic role of minimal residual disease detection in follicular lymphoma.

The identification of patients at high risk of relapse is a critical goal of modern translational research in oncohematology. Minimal residual disease (MRD) detection by polymerase chain reaction-based methods is routinely employed in the management of patients with acute lymphoblastic leukemia. Current knowledge indicates that it is also a useful prognostic tool in several mature lymphoproliferative disorders and particularly in follicular lymphoma (FL). Based on this evidence clinical trials employing MRD-based risk stratification are currently ongoing in FL. In this review the 'state of the art' of MRD evaluation in FL is discussed. A short description of technical issues and recent methodological advances is provided. Then, the bulk of the review focuses on critical take-home messages for clinicians working in the field. Finally, we discuss future perspectives of MRD detection and more generally outcome prediction in FL.

Telomere loss in Philadelphia-negative hematopoiesis after successful treatment of chronic myeloid leukemia: evidence for premature aging of the myeloid compartment.

Telomere shortening, a well-known marker of aging and cellular stress, occurs under several conditions in the hematopoietic compartment, including aplastic anemia and following iatrogenic noxae. We decided to verify whether pathological telomere erosion also arises in restored Philadelphia-negative (Ph-negative) hematopoiesis following successful treatment of chronic myeloid leukemia (CML). Eighty-one CML patients in complete cytogenetic remission were compared to 76 age-matched healthy subjects. Myeloid cells of CML patients had shorter telomeres than controls (6521 bp vs 7233 bp, p<0.001). This difference was specific for the myeloid compartment, since it was not observed in lymphoid cells (6774 bp vs 6909 bp, p=0.620). Acquired Ph-negative cytogenetic abnormalities (p=0.010), lack of complete molecular remission (p=0.016) and age (p=0.013) were independent predictors of telomere shortening. Telomere dynamics were assessed over a median follow-up period of 22 months. We documented accelerated non-physiological ongoing telomere shortening in 17/59 CML patients (28%). Patients experiencing grade 2-4 hematological toxicity, during CML remission possessed significantly shorter telomeres compared to those lacking toxicity (p=0.005 for any toxicity, p=0.007 for anemia). CML patients suffer from significant and often ongoing telomere stress resulting in premature and selective aging of the myeloid compartment which might have long-term consequences on function and integrity of Ph-negative hematopoiesis.

IGF-I induces caveolin 1 tyrosine phosphorylation and translocation in the lipid rafts.

Caveolin 1, a component of caveolae, regulates signalling pathways compartmentalization interacting with tyrosine kinase receptors and their substrates. The role of caveolin 1 in the Insulin Receptor (IR) signalling has been well investigated. On the contrary, the functional link between caveolin 1 and IGF-I Receptor (IGF-IR) remains largely unknown. Here we show that (1) IGF-IR colocalizes with caveolin 1 in the lipid rafts enriched fractions on plasmamembrane in R-IGF-IR(WT) cells, (2) IGF-I induces caveolin 1 phosphorylation at the level of tyrosine 14, (3) this effect is rapid and results in the translocation of caveolin 1 and in the formation of membrane patches on cell surface. These actions are IGF-I specific since we did not detect caveolin 1 redistribution in insulin stimulated R(-) cells overexpressing IRs.