OUP user menu

A Comparative Analysis of Molecular Genetic and Conventional Cytogenetic Detection of Diagnostically Important Translocations in More Than 400 Cases of Acute Leukemia, Highlighting the Frequency of False-Negative Conventional Cytogenetics

Rebecca L. King MD, Mojdeh Naghashpour MD, PhD, Christopher D. Watt MD, PhD, Jennifer J.D. Morrissette PhD, Adam Bagg MD
DOI: http://dx.doi.org/10.1309/AJCPJCW6BY0CNIHD 921-928 First published online: 1 June 2011

Abstract

In this study, we correlated the results of concurrent molecular and cytogenetic detection of entity-defining translocations in adults with acute leukemia to determine the frequency of cryptic translocations missed by conventional cytogenetics (CC) and of recurrent, prognostically relevant translocations not detectable by multiplex reverse transcriptase–polymerase chain reaction (MRP). During a 5.5-year period, 442 diagnostic acute leukemia specimens were submitted for MRP-based detection of 7 common recurrent translocations: t(8;21), t(15;17), inv(16), t(9;22), t(12;21), t(4;11), and t(1;19), with a detection rate of 15.2% (67/442). CC was performed in 330 (74.7%) of 442 cases. In 7 of these 330 cases, CC missed the translocation detected by MRP. In 50 additional cases, CC revealed 1 of the MRP-detectable translocations (all were also MRP positive), yielding a false-negative rate of 12% (7/57) for the CC assay. The remaining 140 of 190 cases with clonal cytogenetic changes harbored abnormalities that were not targeted by the MRP assay, including 8 that define specific acute myeloid leukemia entities. This study revealed the frequent occurrence of false-negative, entity-defining CC analysis and highlighted the complementary nature of MRP and CC approaches in detecting genetic abnormalities in acute leukemia.

Key Words:
  • Acute leukemia
  • Molecular
  • Reverse transcriptase–polymerase chain reaction
  • RT-PCR
  • Cytogenetics
  • Multiplex
  • False-negative
  • Translocations
  • Complementary

The contemporary diagnosis of acute leukemia (AL) relies on a multifaceted approach using morphologic, cytochemical, immunophenotypic, and cytogenetic and molecular analysis.1 Although each of these modalities provides critical information with regard to diagnosis, it has been well established in recent years that genetic studies provide the most disease-defining, prognostically relevant, and therapy-determining set of data.114 ALs harbor a variety of recurrent genetic aberrations, some of which can be ascertained by conventional cytogenetic (CC) analysis.27,1416 These comprise 3 broad categories of abnormalities: (1) balanced chromosomal translocations, typically, but not always, without a net gain or loss of genetic material; (2) numeric abnormalities, such as deletions and additions of whole or segments of chromosomes; and (3) submicroscopic genetic abnormalities. A subset of balanced translocations in ALs are diagnostic of specific entities and have independent prognostic value, which directly impacts therapeutic decision making.1,3,4

Although many of the translocations can be detected (and indeed were discovered) by CC, newer molecular and fluorescence in situ hybridization (FISH) methods for detection of specific abnormalities are becoming routine in the clinical laboratory.17 Molecular methods, specifically reverse transcriptase–polymerase chain reaction (RT-PCR)–based assays, have numerous technical advantages over CC, including shorter turnaround time and no requirement for dividing cells. In addition, prognostically significant translocations may go undetected by CC if they involve regions with similar banding patterns, so-called cryptic translocations.16,1824 Thus, RT-PCR and FISH also have the benefit of detecting aberrations that may be missed by CC analysis.

Studies have demonstrated the excellent diagnostic sensitivity of FISH for detecting translocations and numeric aberrations in AL and have encouraged its use as an adjunct to CC.25,26 FISH, as well as RT-PCR, can be multiplexed, allowing for the investigation of several chromosomal aberrations at once.2528 Multiplex FISH also provides the unique benefit of clarifying complex karyotypes with identification of derivative chromosomes, ring chromosomes, and complex translocations involving more than 2 chromosomes that may not be apparent by CC.2528 However, multiplex FISH is not currently used in routine diagnostic laboratories.

RT-PCR also provides a molecular fingerprint that can be precisely, sensitively, and quantitatively measured, allowing for the subsequent posttherapeutic detection of minimal residual disease.16,19,29,30 Submicroscopic abnormalities, such as point and length mutations, are assuming increased relevance in AL31 and, by definition, can be detected only by molecular approaches. However, despite these obvious advantages, molecular approaches, including RT-PCR–based assays, have not obviated the role for CC (or FISH) studies in the diagnosis and characterization of AL. In addition to detecting the common, relevant translocations, CC has the advantage of identifying numeric aberrations and rare translocations that are not yet well characterized or readily amenable to molecular detection. Indeed, some of these lesions, such as t(1;22), t(6;9), and inv(3), have been incorporated in the World Health Organization (WHO) 2008 classification of hematopoietic malignancies.1

The validation and initial application to routine molecular diagnostic practice of the detection of common prognostically relevant leukemia-associated translocations using multiplex RT-PCR (MRP) have been reported.32,33 The goal of the present study was to detail our experience in a larger cohort of patients during an extended period to evaluate the relative roles of MRP and CC in the classification of AL. In doing so, we sought to establish the usefulness of each method for detecting diagnostically and prognostically relevant translocations. Finally, we wanted to determine the frequency of cryptic translocations missed by CC and of recurrent, prognostically relevant translocations not detectable by the MRP assay.

Materials and Methods

Diagnostic Samples and Study Population

Data were collected on adult patients who had diagnostic specimens submitted for MRP at the Hospital of the University of Pennsylvania, Philadelphia, during a 5.5-year period (July 2004 through December 2009). Patients were excluded if they were ultimately given a diagnosis other than AL. The specimens consisted of peripheral blood or bone marrow aspirate collected into EDTA for MRP and sodium heparin for CC. Diagnoses were established by conventional morphologic, histologic, cytochemical, immunophenotypic, and genetic criteria. The institutional review board of the University of Pennsylvania approved the study.

RNA Isolation

RNA was extracted with a silica gel–based membrane system, the RNAeasy Mini Kit (Qiagen, Valencia, CA) or the QIAmp RNA Blood Mini kit (Qiagen) according to the manufacturer’s specifications and following the recommended protocol adjustments for fresh samples.

Multiplex RT-PCR

MRP was performed by 1 of 2 commercially available assays used at our institution during the study period, Hemavision-7 (HV-7; DNA Technology A/S, Aarhus, Denmark) and Signature LTx (LTx; Asuragen, Austin, TX). Both assays detect 7 frequent translocations/inversions in AL, including the most commonly occurring variants thereof. These are the t(1;19), t(12;21), inv(16), t(8;21), t(4;11), t(15;17), and t(9;22) Table 1.

Hemavision-7

The HV-7 system was used from July 2004 through April 2006. The system was used with the reagents, concentrations, and volumes indicated in the manufacturer’s specifications, as previously published.32 In contrast with the LTx method, the HV-7 uses an agarose gel–based readout. In this system, a specimen was scored positive if a distinct, sharp, and appropriately sized band was detected. Owing to difficulty obtaining reagents, this method was discontinued in our laboratory as of May 2006.

Signature LTx

The LTx assay was used between May 2006 and December 2009. The system was used with the reagents, concentrations, and volumes indicated in the manufacturer’s specifications. In brief, complementary DNA was synthesized using approximately 400 ng of RNA (3 μL), 4 μL of Signature LTx RT Primer Mix (Asuragen), and 3 μL of Signature LTx Diluent (Asuragen). This mixture was then incubated in a thermocycler at 70°C for 5 minutes and then cooled to 4°C. A master mix consisting of 4 μL of Signature LTx Buffer I (Asuragen), 4 μL of Signature LTx Buffer II (Asuragen), 1 μL of MMLV reverse transcriptase (Asuragen), and 1 μL of Signature LTx RNAse inhibitor (Asuragen) was added to each sample. Samples were then incubated in a thermocycler as follows: 42°C for 45 minutes and 93°C for 10 minutes and then cooled to 4°C.

View this table:
Table 1

PCR was performed using 5 μL of Signature LTx DNA Primer Mix (Asuragen), 5 μL of Signature LTx DNA Amp Buffer (Asuragen), 0.5 μL of AmpliTAQ Gold (5U/μL; Applied Biosystems, Foster City, CA), 10 μL of Signature LTx Diluent, and 5 μL of complementary DNA product for a total volume of 25 μL per reaction. PCR conditions were as follows: 1 cycle of 95°C for 10 minutes; 45 cycles of 94°C for 30 seconds, 55°C for 30 seconds, and 72°C for 30 seconds; and hold at 4°C.

Hybridization was performed using 45 μL of Signature LTx Bead Mix (Asuragen) with 5 μL of PCR product. Hybridization conditions were as follows: 95°C for 5 minutes and 52°C for 25 minutes.

Following hybridization, the samples were transferred to the Luminex instrument (Luminex, Austin, TX) where 25 μL of 1× Signature LTx Conjugate (Asuragen) was added to the mixture. Detection on the Luminex instrument was then performed according to the manufacturer’s protocol. In this system, a specimen was scored positive for a specific translocation if the mean fluorescence intensity (MFI) was 400 or more. Specimen RNA quality was assessed through the amplification and detection of endogenous glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Samples with a GAPDH MFI of less than 400 and a translocation MFI of less than 400 were reported as inconclusive. In a preliminary study, the validation of the LTx assay was accomplished by comparing its performance characteristics with the HV-7 in our laboratory, which revealed a sensitivity and specificity of 100% for diagnostic specimens.33 The LTx MRP assay can be completed in 6 hours, allowing for a same-day turnaround time for urgent cases.

Cytogenetic Analysis

CC was performed on unstimulated 24-hour cultures of bone marrow or peripheral blood cells. Trypsin/Giemsabanded metaphases were prepared and analyzed using standard techniques of colchicine arrest, hypotonic treatment, and 3:1 of methanol/acetic acid fixation. In most cases, a karyotype was reported based on the analysis of at least 20 metaphases.

Results

MRP was performed on 494 diagnostic patient specimens during the study. Of these, 52 were excluded for a diagnosis other than AL. Of the remaining 442 cases, acute myeloid leukemia (AML) was diagnosed in 375 (84.8%) and acute lymphoblastic leukemia (ALL) in 67 (15.2%). The HV-7 system was used in 107 specimens and the LTx assay in 335.

Of 442 specimens tested using the MRP assay, 67 (15.2%) were positive for 1 of the detectable translocations Table 2 and Figure 1. When results from the 2 assays were compared, 20 (18.7%) of 107 specimens were positive by the HV-7 and 47 (14.0%) of 335 were positive by the LTx (P = .27). No specimens tested were positive for more than 1 translocation by the molecular assay. Of the 67 MRP+ cases, 19 had additional cytogenetic abnormalities.

Of the 442 specimens, 20 (4.5%) had inconclusive results from the MRP assay owing to suboptimal RNA quality Figure 2. Six of these inconclusive cases were studied with the HV-7 system and 14 with the LTx. Of the 20 cases with an inconclusive MRP result, CC studies were available for 14 (4 CC not ordered; 2 CC failed). In 5 cases, there was a normal karyotype, and 9 demonstrated a clonal cytogenetic abnormality. Of these 9 cases, 1 revealed an inv(16) by CC, which is detectable by our MRP assay. The remaining 8 cases had abnormalities that are not targeted by our MRP assay. Thus, at least 1 inconclusive MRP case harbored a potentially detectable translocation that was evident by CC (Figure 2); however, this cannot be considered a true false-negative because the RNA quality was suboptimal.

View this table:
Table 2
Figure 1

Flow chart highlighting cases detected by MRP but missed by CC (7/57 [12%] false-negative by CC). CC, conventional cytogenetics; MRP, multiplex reverse transcriptase–polymerase chain reaction.

Figure 2

Flow chart highlighting results of all cases evaluated in this study. Inconclusive MRP are cases in which there was suboptimal RNA quality that resulted in poor amplification of control targets in the MRP assay. CC, conventional cytogenetics; MRP, multiplex reverse transcriptase–polymerase chain reaction.

CC was successfully performed on 330 (78.2%) of 422 cases with valid MRP results available (Figure 2). In the remaining 92 cases (21.8%), CC failed (n = 32) or was not ordered (n = 60).

Of the 330 cases with evaluable CC, 140 (42.4%) had normal cytogenetics and 190 (57.6%) had clonal cytogenetic abnormalities (Figure 2). MRP was negative in 135 (96.4%) of the 140 cases with normal cytogenetics. However, the remaining 5 cases each had a translocation detected by MRP that was missed by CC (Figure 2) Table 3.

Among the 190 cases with cytogenetic abnormalities, there were 50 with translocations identified that were potentially detectable by MRP. Of these, all 50 were also positive by MRP. The remaining 140 cases with clonal cytogenetic abnormalities had genetic aberrations identified that were not targeted by the MRP assay. However, of these, 2 were positive by MRP, again revealing translocations missed by CC (Figure 2) Table 4. Therefore, in total, there were 57 translocations identified that could have been detected by the MRP and CC assays. Of these, 50 were indeed detected by both assays, whereas 7 were missed by CC, yielding a false-negative CC rate of 12% (7/57) (Figure 1).

From this group of 190 cases with abnormal CC, we found 6 AML specimens that had the inv(3)(q21;q26.2) or t(3;3)(q21;q26.2)/RPN1-EVI1 translocation, 1 with a t(6;9) (p23;q34.3)/DEK-NUP214, and 1 with a t(1;22)(p13;q13)/RBM15-MKL1 translocation. Although 9 patients with AML had a translocation involving the 11q23 locus detected by CC, none had a t(9;11)(p22;q23). These 9 cases comprised 6 with t(11;19)(q23;p13.1) and 1 each with t(6;11)(q27;q23), t(11;22)(q23;q13), and t(1;11)(p32;q23).

MRP but not CC was done in 92 cases, and of this group, 10 had a translocation detected by the MRP assay. In 4 of the cases, there was a t(15;17), and the patients were diagnosed with acute promyelocytic leukemia (APL), 1 was a case of AML with t(8;21), and 1 was AML with inv(16). The other 4 were cases of B-cell ALL, 3 with a t(9;22) and 1 with a t(1;19).

Discussion

Identification of genetic aberrations is required in the contemporary diagnosis of AL. With the publication of the WHO 2008 classification,1 there are now 7 recognized, genetically defined subgroups of AML and 5 subgroups of B lymphoblastic leukemia. In fact, the presence of t(8;21), inv(16)/t(16;16), and t(15;17) can be considered diagnostic of AML, irrespective of the blast count.1 All 3 of these entities confer a favorable prognosis among patients with AML. Cases with inv(16)/t(16;16) and t(8;21), both of which affect the core-binding factor transcription factor complex, are associated with good response to chemotherapy and relatively long complete remissions.10,11 AML with t(15;17)/PML-RARA (also referred to as APL) is a prototype of a genetic aberration dictating specific therapy, given the key role of all-trans retinoic acid (ATRA).12

View this table:
Table 3
View this table:
Table 4

Treatment decision making (including remission induction chemotherapy, alone or with stem cell transplantation, and timing of stem cell transplantation) depends on multiple parameters; however, genetic data are paramount.10,11,34 Failing to detect genetic abnormalities, including the aforementioned fusion transcripts, could potentially lead to inappropriate therapy, for example, treating with transplantation an apparently cytogenetically normal (CN) patient with AML with an unrecognized favorable prognosis, such as a patient with inv(16), or not exposing a patient with a cryptic t(15;17) to the benefits of ATRA.12

In our study, there were 5 cases in which CC revealed a normal karyotype, in which we found 4 balanced translocations by RT-PCR that could be targets for specific therapy, 1 with PML-RARA and 3 with BCR-ABL1 (Table 3). However, in the patient with APL, there was a suboptimal yield of analyzable metaphases (6 instead of 20). In this case, the t(15;17) detected by MRP was later confirmed with FISH. The morphologic picture was that of APL, and flow cytometry revealed the leukemia cells to be negative for HLA-DR and CD34. All 3 cases of B-cell ALL with a t(9;22) translocation missed by CC were positive for CD33. In addition, in the case with the cryptic inv(16) (Table 4), CC revealed a gain of chromosome 22 (which is often associated with this inversion), while the morphologic picture was myelomonoblastic with abnormal eosinophils, and the blasts expressed CD2. In the case of the t(4;11) missed by CC, the karyotype revealed abnormalities of chromosomes 4 and 11, albeit in no more than 6 cells in this composite karyotype, but did not demonstrate the presence of the t(4;11) translocation (Table 4). Morphologic features in the case of the cryptic t(8;21) were consistent with an AML with maturation. In addition, the blasts were noted to have large salmon-pink granules, a morphologic feature often associated with the t(8;21). Thus, the MRP results in these cases were not entirely unexpected, based on the additional findings.

Of patients with evaluable RT-PCR results, CC analysis was attempted in 362 patient samples. CC analysis failed in 32 samples (8.8%) owing to poor quality and/or lack of adequate metaphases. An additional 60 cases did not have CC ordered. Together, these account for the 21.8% of cases that had only MRP results available in our study. Of these, there were 4 cases in which RT-PCR revealed a t(15;17). In 3 of these cases, there were classic hypergranular APL morphologic features, whereas 1 was a microgranular variant. In one case, FISH confirmed the translocation, and in another, the immunophenotype was compatible with APL. In these 4 cases, it is unclear whether the decision not to order CC analysis was made by the clinical team before or following the reporting of the positive t(15;17) result obtained by MRP. It is likely, however, that the positive t(15;17) by MRP led the clinical team to make therapeutic decisions and opt not to order CC.

In our study, in cases with an optimal RNA yield, the MRP assay showed a diagnostic sensitivity of 100%, being able to detect translocations in all 50 of 50 cases in which they were identified by CC. By contrast, of the 57 cases that harbored 1 of the 7 translocations identified by MRP or CC and that could be directly compared, 7 (12%) had cryptic fusion transcripts that were detected by MRP only but missed by CC (Figure 1 and Tables 3 and 4). This frequency of false-negative CC results is somewhat higher than in other studies in which cryptic fusion transcripts were seen in 0.5% to 10% of patient samples.15,18,3538 The reason for this is unclear, although 1 study demonstrated cryptic fusion transcripts in 15% of cases; however, here, RT-PCR included numerous translocations not detected by our assay, perhaps highlighting the need for more comprehensive MRP studies.39

A targeted review of the 7 cases with cryptic fusion transcripts identified 2 patients in whom therapy had been initiated before the submission of specimens for MRP and CC analysis (Table 3). One was a patient with AML with t(15;17)/PML-RARA who had already been treated with 2 doses of ATRA more than 2 days before sample analysis, whereas another was a patient with ALL with t(9;22)/BCR-ABL1 who had been treated with 1 dose of methotrexate at an outside institution. In both cases, it is likely that the therapy contributed to the false-negative cytogenetic results and that these translocations may not have been truly cryptic. The t(15;17)/PML-RARA case was the same case alluded to earlier in which there was a suboptimal yield of metaphases (6 instead of 20). However, these specimens were legitimately submitted for clinical diagnostic purposes, albeit not under ideal circumstances. Accordingly, although it might be argued that these are not bona fide false-negatives, they nevertheless bolster the value of MRP, given its ability to detect diagnostically relevant genetic events despite the putative negative effects of prior therapy on CC analysis.

Although this study and others highlight the usefulness of molecular (RT-PCR and/or FISH) assays that are rapidly and reliably performed in urgent (eg, APL) and nonemergency situations, the availability of molecular testing should not overshadow the critical role that CC can fulfill in the diagnostic setting. CC analysis, although time-consuming relative to molecular analysis by RT-PCR, reveals global chromosomal abnormalities that are not readily discernible by targeted molecular methods, which have the limited capacity to detect only what they are designed to detect. The importance of analyzing the comprehensive karyotype in AL is highlighted by the recent additions to the list of ALs defined by recurrent translocations or numeric abnormalities in the WHO 2008 classification,1 all of which were initially identified by CC, as they were in the WHO 2001 classification. Furthermore, CC analysis may well lead to the continuous discovery of new, relevant genetic events in leukemia, despite our nascent molecular centricity.

In our study, without CC analysis, we would have missed 6 cases of AML with inv(3)(q21q26.2) or t(3;3)(q21q26.2)/RPN1-EVI1 translocation, which has been shown to be an extremely aggressive disease.40 We also identified 1 case with a t(1;22)(p13q13), RBM15-MKL1, a rare translocation seen in fewer than 1% of AML cases that is associated with megakaryoblastic features.41,42 This case was diagnosed as AML without maturation with nonspecific morphologic features (small blasts with scant basophilic cytoplasm and round nuclei); megakaryoblastic markers (CD41, CD42, CD61, and von Willebrand factor) were not evaluated by flow cytometry or immunohistochemical analysis. In another case of AML, we identified a t(6;9)(p23;q34)/DEK-NUP214, an abnormality associated with bone marrow basophilia and multilineage dysplasia and a poor prognosis.43 We did not encounter any cases of AML t(9;11), which confers an intermediate prognosis and would not have been detectable by our RT-PCR assay. Curiously, we identified 9 AMLs with 11q23/MLL translocations, involving partners other than MLLT3, which is involved in the entity defining t(9;11).

It is also now well established that there are a number of specific gene mutations that afford prognostic significance in AL. This is especially true for AML, in which 40% to 49% of patients are CN, a designation that historically held an “intermediate” prognosis.1,44 It is now recognized that a number of the CN patients (as well as many with recurrent translocations) have submicroscopic gene mutations that impart additional prognostic information.1,44 Examples include mutations in the NPM1 gene, reported in 46% to 62% of CN patients with AML, which affords a relatively good prognosis, and internal tandem duplication of the FLT3 gene (28%–33% of CN patients), which predicts a poor outcome.44 In ALL, recent studies have shown recurrent mutations in genes involved in B-cell differentiation, including PAX5 and IKZF, and these may have prognostic significance as well.44,45 Detection of these mutations, as well as numerous others, relies on molecular methods.

There are a number of limitations in our study. First, the MRP assay is not comprehensive, and it should be updated to incorporate as many WHO 2008 genetically designated entities as possible. Future multiplex assays might also include the ability to test for prognostically relevant mutations alluded to earlier. In addition, the assay includes AML- and ALL-associated translocations in a single reaction; it would be more logical to group AML and ALL translocations separately, to preclude unnecessary testing. Second, our patient population is restricted to adults, with an expected predominance of AML (85% of our cases). However, while acknowledging that the epidemiology of AL in the pediatric population is very different from that seen in adults, we believe that the usefulness and complementarity of these 2 methods would likely hold in a similar study performed in the pediatric population.

This study has demonstrated that molecular and CC analyses for translocations are critical in determining diagnosis and prognosis and in guiding therapy in AL. This is highlighted by the occurrence of cryptic translocations that are missed by CC (false-negatives) but detected by MRP. However, although molecular technology has emerged as the more sensitive (diagnostically and analytically) and facile modality for genetic analysis, it is clear that CC continues to fill an important niche in the diagnosis of AL. Thus, it is important to appreciate the pitfalls of both strategies, and we believe that these methods should be used routinely in a complementary manner in clinical practice.

CME/SAM

Upon completion of this activity you will be able to:

  • discuss the significance of recurrent genetic aberrations in acute leukemia and the methods that are available for their detection.

  • compare the utility of conventional cytogenetics with that of multiplex reverse transcriptase–polymerase chain reaction in the classification of acute leukemia.

The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this educational activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.

The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Questions appear on p 976. Exam is located at www.ascp.org/ajcpcme.

Footnotes

  • Supported by a SCOR grant from the Leukemia and Lymphoma Society of America, White Plains, NY (Dr Bagg); and by a grant from the J.P. McCarthy Fund (Dr Bagg).

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
View Abstract