ALBERT

Description: The diagnosis of myelodysplastic syndrome (MDS) currently relies primarily on the morphologic assessment of the patient's bone marrow and peripheral blood cells. Moreover, prognostic scoring systems rely on observer-dependent assessments of blast percentage and dysplasia. Gene expression profiling could enhance current diagnostic and prognostic systems by providing a set of standardized, objective gene signatures. Within the Microarray Innovations in LEukemia study, a diagnostic classification model was investigated to distinguish the distinct subclasses of pediatric and adult leukemia, as well as MDS. Overall, the accuracy of the diagnostic classification model for subtyping leukemia was approximately 93%, but this was not reflected for the MDS samples giving only approximately 50% accuracy. Discordant samples of MDS were classified either into acute myeloid leukemia (AML) or “none-of-the-targets” (neither leukemia nor MDS) categories. To clarify the discordant results, all submitted 174 MDS samples were externally reviewed, although this did not improve the molecular classification results. However, a significant correlation was noted between the AML and “none-of-the-targets” categories and prognosis, leading to a prognostic classification model to predict for time-dependent probability of leukemic transformation. The prognostic classification model accurately discriminated patients with a rapid transformation to AML within 18 months from those with more indolent disease.

Description: Gene expression profiling has the potential to offer consistent objective diagnostic test results once a standardized protocol is established. We investigated the robustness, precision, and reproducibility of this technology and present data that complements the Microarray Innovations in LEukemia study (MILE study). In four laboratories, located in Germany (D), Austria (A), and Switzerland (CH) (DACH study), replicates of 112 patient samples were analyzed using the AmpliChip Leukemia research test. Patient samples were centrally collected and diagnosed in daily routine at the Munich Leukemia Laboratory and represented 8 distinct classes of acute and chronic leukemias, with non-leukemia as control group. After purification of the mononuclear cells by Ficoll density centrifugation, 4 × 5 million cells were frozen in lysis buffer and stored at −80°C. Equipped with identical instruments, software, and reagents, study operators were trained on the microarray sample preparation protocol using total RNA from commercially available cell lines. Upon receipt of the frozen lysates each of the four laboratories purified the total RNA from the 112 technical quadruplicates. 99.3% (445/448) of the sample preparations were successfully performed. On average, 8.4 μg, 7.2 μg, 7.4 μg, or 7.5 μg of total RNA, respectively, were isolated from the mononuclear cells from the four laboratories. In three samples less than 1.0 μg of total RNA was obtained and thus the preparation failed. Bland-Altman plots of agreement showed that any two centers were unlikely to have more than an 8.3 μg difference in yield of total RNA from the same sample. On average there was between 0.1 μg to 1.2 μg difference in total RNA yield from the same sample. Further processing of the 445 samples resulted in 437 (98.2%) successfully performed in vitro transcription reactions, i.e. obtained cRNA yield of 〉8.0 μg. On average there was between 0.4 μg to 7.4 μg difference in cRNA yield from the same sample. After hybridization to microarrays on average, 46.1%, 48.6%, 46.5%, and 47.3% of probe sets were detected as present with mean scaling factors of 4.3, 2.9, 3.9, and 3.7, respectively. The mean values and standard deviations of distributions of the coefficient of variation (CV) within each site over all the probe sets of the quantile normalized signals on the chip were 27.2% (StdDev: 12.3%), 27.0% (StdDev: 12.3%), 27.3% (StdDev: 12.3%), 26.9% (StdDev: 12.4%), respectively. Furthermore, in unsupervised hierarchical cluster and principal component analyses replicates from the same patient always clustered closely together, with no indications of association between gene expression profiles due to different operators or laboratories. In conclusion, we demonstrated that microarray analysis can be performed with remarkably high inter-laboratory reproducibility and with comparable quality and high technical precision across laboratories.

Description: Microarray analysis can identify differentially expressed genes associated with distinct clinical and therapeutically relevant classes of both pediatric and adult leukemias. Recently, the MILE (Microarray Innovations in Leukemia) research study program has been launched in 10 centers: 7 from the European Leukemia Network (ELN, WP13) and 3 from the US. In this study, which will include 4,000 patients, the clinical accuracy of gene expression profiles of 16 acute and chronic leukemia subclasses, MDS, and non-leukemia as control will be assessed as compared to current routine diagnostic workup. Each center is trained on an identical microarray protocol and uses the same laboratory equipment, kits, and reagents for target preparation (Affymetrix HG-U133 Plus 2.0). First, the intra- and inter-laboratory comparability was investigated using 2 different cell line samples, MCF-7 and HEPG2, with different amounts of starting material (1 μg and 5 μg input for cDNA synthesis). Also, each center prepared in parallel total RNA and processed replicate samples from three leukemia pts (AML with t(8;21), CML, and CLL). We found a high reproducibility among the different centers: unsupervised analyses accordingly group the two different cell lines distinct from the three types of leukemia samples. In hierarchical clustering and principal component analysis the non-leukemia samples are clearly distinct from the leukemia samples and no clustering of the individual centers can be seen. Remarkably, for the replicates of the leukemia samples the squared correlation coefficients of gene expression range between 0.975 and 0.997 for CML, between 0.975 and 0.998 for CLL, and between 0.970 and 0.999 for the AML with t(8;21). Secondly, the samples were analyzed by a classification algorithm. The algorithm was trained on a database that contains gene expression profiles of 〉1,600 leukemia patients and cell lines and can distinguish 16 different classes of leukemia, MDS, and non-leukemia. Several methods are used to form linear classifiers for all 18 * (18 – 1)/2 = 153 class pairs. The average cross-validation accuracy is 91% or higher. Miscalls are predominantly seen in the distinction between MDS and AML with normal karyotype. The accuracy of resubstitution (application of the classifier to the data forming the classifier) is 100%. For the new data accurate predictions for the non-leukemia cell lines, AML with t(8;21), and CLL were observed. Interestingly, the CML in blast crisis is predicted as AML with other abnormalities. This may be due to the fact that the classifier was trained on CML in chronic phase only. In conclusion, for the first time an international multi-center research study demonstrates a very high reproducibility of microarray analyses performed at different centers for the same leukemia samples. This lays the foundation for an international clinical research initiative evaluating the application of microarrays in the diagnosis and classification of hematological malignancies.

Description: During the years of 2005 to 2008, the MILE (Microarray Innovations in LEukemia) study research program was performed in 11 laboratories across three continents: 7 from the European Leukemia Network (ELN, WP13), 3 from the US and 1 in Singapore. The first stage was designed as biomarker discovery phase to generate whole-genome gene expression profiles (GEP) from recognized categories of clinically relevant leukemias and myelodysplastic syndromes (MDS). These were C1: mature B-ALL with t(8;14), C2: pro-B-ALL with t(11q23)/MLL, C3: c-ALL/pre-B-ALL with t(9;22), C4: T-ALL, C5: ALL with t(12;21), C6: ALL with t(1;19), C7: ALL with hyperdiploid karyotype, C8: c-ALL/pre-B-ALL without specific genetic abnormalities, C9: AML with t(8;21), C10: AML with t(15;17), C11: AML with inv(16)/t(16;16), C12: AML with t(11q23)/MLL, C13: AML with normal karyotype or other abnormalities, C14: AML with complex aberrant karyotype, C15: CLL, C16: CML, C17: MDS, and C18: non-leukemic and healthy bone marrow samples as controls and were compared to conventional diagnostic assays (“Gold Standard”). Data from the completed MILE Stage I included 2143 retrospectively collected adult and pediatric samples tested with HG-U133 Plus 2.0 microarrays (Affymetrix). In total only 47 analyses (2.1%) failed technical quality criteria. Cross-validation accuracy (average of three 30-fold cross-validations) of the final 2096 MILE Stage I samples was 92.1% concordant with the center-specific “Gold Standard” diagnosis (average call rate 99.4%). In nine classes the sensitivity was ≥94.3%: C2, C3, C4, C5, C9, C10, C11, C15, and C16. Lower sensitivities were observed for C7, C8, C14, and C17; which can largely be explained by the biological heterogeneity and non-standardized “Gold Standard” definitions for these entities. Yet, it is notable that all these classes showed specificities above 98.1%. In order to assess the clinical utility of microarray-based diagnostics a prospective Stage II was subsequently performed using a customized microarray representing 1480 probe sets. Overall, 1156 high quality GEP have been generated in MILE Stage II and represent an independent and blinded test set for the algorithms developed. A focused classification scheme aimed at accurately addressing only acute leukemias resulted in a 95.5% median sensitivity and a 99.5% median specificity for the 14 classes included in the classifier (C1 – C14, n=696). Lower accuracies were observed for the interface of C7–C8 in ALL, as well as C12 and C14 in AML. Interestingly, during the process of discrepant results analyses, it was observed that for 7.5% (n=52) of acute leukemias microarray results were correctly diagnosing samples as compared to the initial “Gold Standard” diagnoses entered into the study database, either because of erroneous entries into case report forms (24%) or subsequent re-testing of left-over material following the suggested diagnosis from the microarray (76%). In addition, predicted accuracies for CLL, CML and MDS in Stage II were 99.2%, 95.2%, and 81.5%, respectively. In conclusion, the MILE research study confirms in a final cohort of 3252 patients that microarrays accurately classify acute and chronic leukemia samples into known diagnostic and prognostic sub-categories. This final report underlines that the standardized method of gene expression profiling with low technical failure rate and simplified standard operating procedures may improve current “Gold Standards” as an adjunct to conventional diagnostic algorithms and potentially offers a reliable diagnostic/prognostic tool for many patients who don’t have access to a state-of-the-art “Gold Standard” workup. Our gene expression database, intended to be submitted to the public domain, will further contribute to research that aims to elucidate the molecular understanding of leukemias.

Description: The MILE (Microarray Innovations in LEukemia) study has previously shown that gene expression signatures associated with initial leukaemia classifier (LCver7) give an overall cross-validation accuracy of 〉95% for distinct sub-classes of pediatric and adult leukemias. However, only 50% of the 174 MDS samples in the whole-genome microarray analysis (Stage 1) of the MILE study were correctly identified; the remainder showed AML-like or non-leukemia-like gene profiles. An external morphological review (DB & HL) according to FAB and WHO criteria, of the 174 slides was performed independently (blind) which resulted in 6 samples being reclassified as AML and 4 non-leukemia cases excluded from the study. A recently improved, hierarchical based algorithm correctly identified 100% of the confirmed MDS cases. In this study, using LCver7, the confirmed 164 samples had 50% MDS classifications (Class 17), 23.8% non-leukemia classifications (Class 18), and 22.6% AML classifications (Classes 13 or 14) with the remaining 3.7% having a classification tie between 2 or 3 Classes (due to low confidence). No 5q- syndrome patients had an AML call, whilst 68.3% of RAEB2 patients had an AML classification and none were Class 18. Similarly, 95.6% of Low IPSS patients were classified as Class 17 or 18, whilst all patients (n=5) with High IPSS had an AML call. The classification was independent of blast cells: 10.2% of Class 18 calls had 〉5% blasts; 28.2% of AML-like cases had

Description: Key Points CD30 expression defines a novel and unique subgroup of DLBCL with favorable clinical outcome and distinct gene expression signature.

Description: Key Points DLBCL patients with MYC/BCL2 coexpression demonstrate inferior prognosis and high-risk gene expression signatures.

Description: Gene expression microarrays had been used to classify known tumor types and various hematological malignancies (Yeoh et al, Cancer Cell 2002; Kohlmann et al, Genes Chromosomes Cancer 2003), enforcing the objective that microarray analysis could be introduced soon in the routine classification of cancer (Haferlach et al, Blood 2005). However, there’re still doubts about gene expression experiments performance in clinical laboratory diagnosis. For instance, the quality of starting material is a major concern in microarray technology and there are no data on the variation in gene expression profiles ensuing from different RNA extraction procedures. Here, as part of the internal multicenter MILE Study program, we assess the impact of different RNA preparation methods on gene expression data, analyzing 27 patients representative of nine different subtypes of pediatric acute leukemias. We compared the three currently most used protocols to isolate RNA for routine diagnosis (PCR assays) and microarray experiments. They are named as method A: lysis of mononuclear leukemia cells, followed by lysate homogeniziation, followed by total RNA isolation; method B: TRIzol RNA isolation, and method C: TRIzol RNA isolation followed by total RNA purification step. The methods were analyzed in triplicates for each sample (24) and additional three samples were performed in technical replicates of three data sets for each preparation (HG-U133 Plus 2.0). Method A results in better total RNA quality as demonstrated by 3′/5′ GAPD ratios and by RNA degradation plots. High comparability of gene expression data is found between samples in the same leukemia subclasses and collected with different RNA preparation methods thus demonstrating that sample preparation procedures do not impair the overall signal distribution. Unsupervised analyses showed clustering of samples first by each patient’s replicate conditions, then by leukemia type, and finally by leukemia lineage. In fact, B-ALL samples are clustered together, separately from T-ALL and AML, demonstrating that clustering reflects biological differences between leukemias and that the RNA isolation method is a secondary effect. Also, supervised cluster analyses highlight that samples are grouped depending on intra-lineage features (i.e. chromosomal aberrations) thus confirming the clustering organizations as reported in recent gene expression profiling studies of acute leukemias. Our study shows that biological features of pediatric acute leukemia classes largely exceed the variations between different total RNA sample preparation protocols. However, technical replicates analyses reveal that gene expression data from method A have the lowest degree of variation, are more reproducible and more precise as compared to the other two methods. Furthermore, compared to methods B and C, method A produces more differentially expressed probe sets between distinct leukemia classes and is therefore considered the more robust RNA isolation procedure for gene expression experiments using high-density microarray technology. We therefore conclude that method A (initial homogenization of the leukemia cell lysate followed by total RNA isolation) combined with a standardized microarray analysis protocol is highly reproducible and contributes to robustness of gene expression data and that this procedure is most practical for a routine laboratory use.

Description: Gene expression profiling (GEP) is a powerful technology for the molecular analysis of leukemia and it groups biologically defined disease entities into distinct sub-classes that can provide diagnosis, guide therapy, and even correlate with disease prognosis. The experimental procedures of micorarray analysis are often cumbersome and provide ample opportunity for variability in gene expression data. We previously reported on our efforts to standardize micorarray analysis across 11 participating laboratories within the international MILE study (Microarray Innovations in LEukemia) where a large dataset of over 4,000 leukemia patient samples is being generated using both Affymetrix HG-U133 Plus 2.0 and custom format microarrays. For a better applicability in a routine laboratory workflow and in order to improve the robustness of the micorarray analysis we now have modified the original micorarray sample preparation protocols as originally published by the manufacturer. Here we report the final results of this effort to minimize the complexity of the sample preparation protocol and to reduce the time that is necessary to run the assay. We designed pre-assembled kits for total RNA preparation, nucleic acid cleanup, cDNA synthesis, in vitro transcription, hybridization and staining, and wash buffers guiding the operator through the whole process of sample preparation to microarray result generation. To further improve the ease of use of this assay we minimized to a large extent the overall complexity of sample amplification and labeling, as well as target hybridization and detection procedures. For example, for the RNA amplification, cRNA labeling, and signal detection process, the number of individual reagent vials was reduced from 32 to 13 vials. This was achieved by combining individual components to ready-to-use master mixes. Furthermore, starting from total RNA, the time required for generation of labeled and fragmented cRNA has been reduced to a convenient eight hour work-shift. Overall, compared to the original 48 hour protocol as recommended by the manufacturer the new workflow generates microarray data in 26 hours. In total, this development program included n=900 whole genome microarray tests. By comparison testing of the original and the final modified protocols on we further can demonstrate by squared correlation coefficients both high inter-assay (R2 〉 0.9) and intra-assay (R2 〉 0.9) reproducibility and precision of gene expression data of this new sample preparation method. Data from cell lines, normal bone marrow, as well as leukemia samples representing the subclasses AML with normal karyotype or other abnormalities, AML with complex aberrant karyotype, CML, and CLL indicate that reproducible subclassification of leukemias is feasible as all samples were predicted by a classification algorithm as the same class as when the samples were prepared according to the the original method. In conclusion, we developed a robust sample processing methodology for microarray analysis of leukemia samples that allows to generate standardized and reproducible microarray results in multiple laboratories.

Description: TP53 mutation is an independent marker of poor prognosis in patients with diffuse large B-cell lymphoma (DLBCL) treated with cyclophosphamide, hydroxydaunorubicin, vincristine, and prednisone (CHOP) therapy. However, its prognostic value in the rituximab immunochemotherapy era remains undefined. In the present study of a large cohort of DLBCL patients treated with rituximab plus CHOP (R-CHOP), we show that those with TP53 mutations had worse overall and progression-free survival compared with those without. Unlike earlier studies of patients treated with CHOP, TP53 mutation has predictive value for R-CHOP–treated patients with either the germinal center B-cell or activated B-cell DLBCL subtypes. Furthermore, we identified the loop-sheet-helix and L3 motifs in the DNA-binding domain to be the most critical structures for maintaining p53 function. In contrast, TP53 deletion and loss of heterozygosity did not confer worse survival. If gene mutation data are not available, immunohistochemical analysis showing 〉 50% cells expressing p53 protein is a useful surrogate and was able to stratify patients with significantly different prognoses. We conclude that assessment of TP53 mutation status is important for stratifying R-CHOP–treated patients into distinct prognostic subsets and has significant value in the design of future therapeutic strategies.