In acute lymphoblastic leukemia, flow cytometry detects more accurately leukemic cells in patients' cerebrospinal fluid compared to conventional cytology. However, the clinical significance of flow cytometry positivity with a negative cytology - occult central nervous system disease - is not clear. In the framework of the national Campus ALL program, we retrospectively evaluated the incidence of occult central nervous system disease and its impact on outcome in 240 adult patients with newly diagnosed acute lymphoblastic leukemia. All cerebrospinal fluid samples were investigated by conventional cytology and flow cytometry. The presence of ≥10 phenotypically abnormal events, forming a cluster, was considered as flow cytometry positivity. No central nervous system involvement was documented in 179 patients, while 18 were positive by conventional morphology and 43 were occult central nervous system disease positive. The relapse rate was significantly lower in central nervous system disease negative patients and the disease-free and overall survival were significantly longer in central nervous system disease negative patients than in those with manifest or occult central nervous system disease positive. In multivariate analysis, the status of manifest and occult central nervous system disease positivity was independently associated with a worse overall survival. In conclusion, we demonstrate that in adult acute lymphoblastic leukemia patients at diagnosis flow cytometry can detect occult central nervous system disease at high sensitivity and that the status of occult central nervous system disease positivity is associated with an adverse outcome. (Clinicaltrials.gov NCT03803670)
Over the last two decades, improved response rates have been reported in adult patients with acute lymphoblastic leukemia (ALL).1-3 In this context of a superior systemic disease control, central nervous system (CNS) involvement has become an ever more influential limitation to the achievement of a long-term cure and a main cause of mortality. At diagnosis, about 5-10% of adult ALL patients have CNS involvement, 4-6 which translates into a shorter overall survival (OS) compared to that of patients without CNS involvement. 4
Conventional cytology (CC) examination of the cerebrospinal fluid (CSF) remains the gold standard for the diagnosis of CNS involvement in ALL; CC is estimated to have a >95% specificity. However, it has a relatively low sensitivity (<50%), resulting in frequent false negative determinations. Such a low sensitivity is due to the poor cellularity of CSF and to the difficulties in distinguishing benign from malignant cells on morphologic grounds only.7,8
Flow cytometric (FCM) immunophenotyping is a valuable tool for the diagnosis and staging of hematologic disorders involving lymph nodes, blood, bone marrow and other body fluids. Current FCM assays allow detection of phenotypically abnormal cells up to the limit of at least 0.01% (1 target cell in 104 events), representing, therefore, a very effective tool for minimal residual disease monitoring in acute leukemia.9 Indeed, several recently published experiences have demonstrated the superior sensitivity of FCM over CC for the detection of CNS disease in patients with ALL and non-Hodgkin lymphoma.10-13 These studies have also contributed to establish a new standard that is the so-called “occult CNS disease” (OCNSD), namely the status of FCM positivity and CC negativity. None of these reports has, however, clarified whether a condition of OCNSD has an additional prognostic role compared to the well-established negative impact of CC positivity. We therefore conducted a multicenter, retrospective study in the framework of the national Campus ALL program aimed at improving the management of adult ALL patients. The aims of the present study were: (i) to evaluate the incidence of OCNSD in a large series of adult patients with ALL; and (ii) to assess the impact of OCNSD on the clinical outcome of these patients.
Study design and patients
Our retrospective analysis included patients seen between January 2007 and December 2017 at 13 Italian hematology centers. Cases were documented using a case report form. Variables included the following data: age, sex, ALL onset, genetic/cytogenetic features, B/T phenotype, white blood cell count (WBC) at diagnosis and at the time of lumbar puncture (LP), lactate dehydrogenase (LDH), chemotherapy, date of complete remission (CR), CSF cell count and chemistry, CC and FCM results, date of systemic and/or CNS relapse, allogeneic stem cell transplant (ASCT), date of death or the last follow-up. Personal information was treated in a confidential manner and all sensitive data were analyzed anonymously. Samples were collected at diagnosis. In patients with a high WBC count, which might be due to the traumatic procedure and confound the CSF picture, the explorative LP was performed once the WBC count was reduced below 10x109/L by administering steroids.
Patients were treated within or according to GIMEMA (LAL0904, LAL1308, LAL1913, LAL1104)14 or NILG (NILGALL10/ 07)15,16 protocols or the Hyper-CVAD/MTX-ARAC regimen. 17,18 In the GIMEMA protocols, CNS prophylaxis consisted in intrathecal injection (IT) of methotrexate (12.5 or 15 mg) alone or combined with steroids once a week for a total of 3-4 administrations during the induction and consolidation cycles, respectively. In LAL0904, cranio-spinal irradiation (CI) was dispensed after the consolidation phase,14 while in the other GIMEMA/NILG protocols CI was omitted and all patients received a CNS-crossing agent-based chemotherapy. According to the NILG-ALL10/07 protocol, 12 triple agent (methotrexate 12.5 mg, cytarabine 50 mg, dexamethasone 4 mg) IT injections were given as CNS prophylaxis. Finally, in the Hyper-CVAD/MTX-ARAC program, 16 prophylactic IT were planned.17,18 CNS therapy for patients with a CCpositive LP consisted of IT injections of 12 mg methotrexate, 50 mg cytarabine and 10 mg methylprednisolone twice weekly until CSF blast clearance, and then once weekly for two administrations.
Cell counts and conventional cytology
Cytospins for CC examination were prepared as previously described in detail.19,20 CC positivity was defined as unequivocal, morphological evidence of leukemic blast in the CSF and/or a CSF WBC count ≥5/L with less than 10 erythrocytes/L.3,21 Traumatic LP were excluded from the analysis.
Flow cytometry analysis
All centers involved were selected on the basis of a strict adherence to a standardized approach relying on the same procedures (time elapsed from collection to processing, number of fluorochromes, number of acquired events and analysis). Samples for FCM analysis were locally processed within 60 minutes from harvest, as described elsewhere.19 A cocktail of 6-8 monoclonal antibodies was used (Online Supplementary Table S1). On average, 1,080 events were acquired (range 0-210,000). In agreement with the recommendations for the analysis of rare events, a cluster of at least 10 phenotypically abnormal events was regarded as proof of CSF infiltration10 (Figure 1). Traumatic LP were excluded from the analysis.
The statistical analysis is described in the Online Supplementary Appendix.
Approval of the local institutional review board and ethics committee was obtained at all participating sites. The trial was registered at clinicaltrials.gov identifier: NCT03803670.
The clinical and laboratory characteristics of the 240 patients are summarized in Table 1. At diagnosis, 179 (75%) CSF samples were negative by both FCM and CC (CNSneg), while 43 (18%) were OCNSD positive (positive by FCM and negative by CC=OCNSDpos) and 18 (7%) were positive by both FCM and CC (manifest CNS disease positive = MCNSDpos) (Table 1). No case proved to be FCM-negative and CC-positive.
The characteristics of patients belonging to the three groups are listed in Table 1. There was an equal male:female ratio among CNSneg, OCNSDpos and MCNSDpos patients. There was no significant difference in median age, median WBC count, B/T lineage, LDH levels between the three patient categories. Cytogenetic/genetic data were available in 178 of 240 cases (74%) and no difference in distribution among the three categories was observed. On the other hand, the status of OCNSDpos and MCNSDpos was significantly associated with a high CSF cellularity (P<0.001) (Table 1) and the levels of CSF proteins (P=0.023) (Table 1). One hundred and seventy-one patients (71%) were treated within or according to GIMEMA protocols, 37 (15%) with the Hyper- CVAD/MTX-ARAC regimen, and 32 (14%) according to the NILG ALL10/07 protocol. Considering the heterogeneity of the chemotherapy regimens utilized, we analyzed our series dividing the patients into three groups on the basis of the intensity of the treatment received. Accordingly, 91 patients (37.9%) underwent a conventional treatment, 120 (50%) an intensified pediatric-inspired regimen, and 29 (12.1%), qualified as unfit or frail, were treated with a reduced intensity schedule (Table 1).
Of the 232 evaluable patients, 198 (85%) achieved a CR with no significant differences between the three CNS status- based groups (P=0.3). Of these 198 patients, 116 (59%) experienced a relapse; in 18 of 116 (15%), disease recurrence occurred in the CNS alone or was combined with a hematologic relapse. The relapse rate was significantly higher in OCNSDpos and MCNSDpos patients than in CNSneg patients (P=0.001) (Table 2). The 3-year disease-free survival (DFS) was also significantly longer in CNSneg patients compared to OCNSDpos or MCNSDpos patients: 39% (95% confidence interval [CI]: 31-48) versus 21% (95%CI: 4.5-33.9) versus 21% (95%CI: 7.9-58.4), respectively (P=0.005) (Table 2). On the contrary, there was no difference in 3- year DFS between OCNSDpos and MCNSDpos patients (P=0.3) (Figure 2).
The 3-year overall survival (OS) in CNSneg, OCNSDpos and MCNSDpos patients was 53% (95%CI: 45.5-61.5), 31% (95%CI: 19.2-50.5) and 22% (95%CI: 9.4-52.7), respectively (P<0.0001) (Table 2). There was no difference in 3-year OS between OCNSDpos and MCNSDpos patients (P=0.2) (Figure 3).
The clinical impact of the CNS status on OS was also challenged in the multivariate Cox proportional hazard analysis applied to models including age, transplant, sex, WBC count and treatment received. Multivariate analysis confirmed that the OCNSDpos (HR=1.82, 95%CI: 1.15-5.92; P=0.01) or MCNSDpos status (HR=3.23, 95%CI: 1.76-2.89; P<0.0001), defined at the time of diagnosis, were factors that independently impacted on OS together with the treatment regimens (intensified vs. conventional vs. reduced intensity for age) (Table 3).
This retrospective study shows that FCM offers better technical support than CC in detecting leukemic cells in the CFS of adult patients with ALL, and documents the clinical impact of OCNSD on the outcome of these patients. By introducing FCM analysis, the detection power improved to such an extent that evidence of CNS involvement increased from 7% to 25% of ALL cases at diagnosis. This analysis confirms previous reports that demonstrated the superior sensitivity of FCM over CC.10,12,13,22,23 In a large retrospective study of 326 CSF samples collected from patients affected by diffuse large B-cell and Burkitt lymphomas, a CSF involvement was detected by FCM in 33 (13%) diffuse large B-cell lymphomas and in 9 (11%) Burkitt lymphomas.24 FCM allows detection of a hematologic disease in CSF specimens even when the cellularity is very low.9,25 This peculiarity has been confirmed in pediatric ALL patients where FCM was able to substantially improve recognition of occult CSF involvement. 26-28 In agreement with pediatric reports,27 the CNS status of our adults did not correlate with risk factors associated with the risk of relapse, such as WBC count at onset, B/T phenotype or cytogenetic/genetic features.
In pediatric ALL, FCM positivity alone in the absence of a positive CC seems to affect clinical outcome.27-29 Similar observations have been made in patients with high-risk non-Hodgkin lymphomas and Burkitt lymphomas, in whom FCM positivity of CSF was associated with a significantly higher risk of CNS relapse and a worse prognosis. 24,30
In adult ALL patients, the role of OCNSD is less clear because of the limited number of studies based on small series of patients. By analyzing 168 CSF samples collected from 31 patients with ALL, Subira et al.31 reported a concordance between FCM and CC except for ten samples. All patients found to be FCM negative remained free from CNS disease. In a small population of 38 adults with ALL or lymphoblastic lymphoma, we previously observed that the median OS of patients with FCM single positivity was intermediate between double positive or negative patients.19
The uncertain clinical significance of the FCM analysis of CSF is confirmed by the discordant position of the current guidelines. In fact, while the National Comprehensive Cancer Network (NCCN) guidelines32 do not indicate that FCM analysis of the CSF should be part of the initial work-up, the more recent American pocket guide for the clinician recommends (although not strongly) performing this examination at diagnosis.33 Based on our large multicenter report, occult CNS status does indeed have a significant impact on outcome. In fact, patients with OCNSD had a worse DFS and OS compared to those who were OCNSD negative. The superimposable duration of OS of OCNSD and MCNSD patients indicates that even the presence of a few cells in the CNS sanctuary has a clinical impact; these few cells can only be detected by using approaches more sensitive than CC, such as FCM. The pronounced neurotropism of ALL34-36 can be responsible for disease recurrence once the leukemic cells, having survived systemic chemotherapy within the CNS sanctuary, migrate to the circulation.37,38 Thus, the availability of highly sensitive methods capable of accurately defining whether or not the CSF is colonized by leukemic cells not only offers a refined diagnostic/prognostic work-up, but also helps to personalize CNS prophylaxis through the early identification of patients who may benefit from more aggressive approaches.
With the limitations of its retrospective nature, the results of our study demonstrate that, in adult ALL patients, FCM can more precisely identify and quantify the number of patients with CNS involvement at diagnosis and that this impacts significantly on the clinical course and outcome of the disease, thus enabling a further refinement of the current diagnostic risk-stratification process. This refined CNS evaluation should become a routine tool for the work-up of ALL patients at presentation. Further and larger prospective studies are needed to further standardize the procedures and promote optimal clinical application of this technique.
- Received July 29, 2019
- Accepted December 20, 2019
- Thomas X, Boiron JM, Huguet F. Outcome of treatment in adults with acute lymphoblastic leukemia: analysis of LALA- 94 trial. J Clin Oncol. 2004; 22(20):4075-4086. https://doi.org/10.1200/JCO.2004.10.050PubMedGoogle Scholar
- Kantarjian HM, O’Brien S, Smith TL. Results of treatment with Hyper-CVAD, a dose-intensive regimen in adult acute lymphocytic leukemia. J Clin Oncol. 2000; 18(3):547-561. Google Scholar
- Thomas X, Le QH. Central nervous system involvement in adult acute lymphoblastic leukemia. Hematology. 2008; 13(5):293-302. https://doi.org/10.1179/102453308X343374PubMedGoogle Scholar
- Lazarus HM, Richards SM, Chopra R. Medical Research Council (MRC)/National Cancer Research Institute (NCRI) Adult Leukaemia Working Party of the United Kingdom and the Eastern Cooperative Oncology Group. Central nervous system involvement in adult acute lymphoblastic leukemia at diagnosis: results from international ALL trial MRC UKALL XII/ECOG E2993. Blood. 2006; 108(2):65-72. https://doi.org/10.1182/blood-2005-11-4666PubMedPubMed CentralGoogle Scholar
- Jabbour E, Thomas D, Cortes J, Kantarjian H, O’Brien S.. Central nervous system prophylaxis in adults with acute lymphoblastic leukemia. Cancer. 2010; 116(10):2290-2300. https://doi.org/10.1002/cncr.25008PubMedGoogle Scholar
- Larson RA. Managing CNS disease in adults with acute lymphoblastic leukemia. Leuk Lymphoma. 2018; 59:3-13. https://doi.org/10.1080/10428194.2017.1326597PubMedGoogle Scholar
- Bromberg JE, Breems DA, Kraan J. CSF flow cytometry greatly improves diagnostic accuracy in CNS hematologic malignancies. Neurology. 2007; 68(20):1674-1679. https://doi.org/10.1212/01.wnl.0000261909.28915.83PubMedGoogle Scholar
- Kaplan JG, DeSouza TG, Farkash A. Leptomeningeal metastases: comparison of clinical features and laboratory data of solid tumors, lymphomas and leukemias. J Neurooncol. 1990; 9(3):225-229. https://doi.org/10.1007/BF02341153PubMedGoogle Scholar
- Craig F, Foon KA. Flow cytometric immunophenotyping for hematologic neoplasms. Blood. 2008; 111(8):3941-3967. https://doi.org/10.1182/blood-2007-11-120535PubMedGoogle Scholar
- Quijano S, Lopez A, Manuel Sancho J. Spanish Group for the Study of CNS disease in NHL. Identification of leptomeningeal disease in aggressive B-cell non Hogkin’s lymphoma: improved sensitivity of flow cytometry. J Clin Oncol. 2009; 27(9):1462-1469. https://doi.org/10.1200/JCO.2008.17.7089PubMedGoogle Scholar
- de Graaf MT, de Jongste AH, Kraan J, Boonstra JG, Sillevis Smitt PA, Gratama JW. Flow cytometric characterization of cerebrospinal fluid cells. Cytometry B Clin Cytom. 2011; 80(5):271-281. https://doi.org/10.1002/cyto.b.20603PubMedGoogle Scholar
- Zeiser R, Burger JA, Bley TA, Windfuhr-Blum M, Schulte-Monting J, Behringer DM. Clinical follow-up indicates differential accuracy of magnetic resonance imaging and immunocytology of the cerebral spinal fluid for the diagnosis of neoplastic meningitis-a single centre experience. Br J Haematol. 2004; 124(6):762-768. https://doi.org/10.1111/j.1365-2141.2004.04853.xPubMedGoogle Scholar
- Di Noto R, Scalia G, Abate G. Critical role of multidimensional flow cytometry in detecting occult leptomeningeal disease in newly diagnosed aggressive B-cell lymphomas. Leuk Res. 2008; 32(8):1196. https://doi.org/10.1016/j.leukres.2007.12.016PubMedGoogle Scholar
- Annino L, Vignetti M, Paoloni FP. Treatment of adolescents and young adults with acute lymphoblastic leukemia (ALL): an update of the GIMEMA experience. Blood. 2009; 114(22):3097. https://doi.org/10.1182/blood.V114.22.3097.3097PubMedGoogle Scholar
- Bassan R, Masciulli A, Intermesoli T. Randomizad trial of radiation-free central nervous system prophylaxis comparing intrathecal triple therapy with liposomal cytarabine in acute lymphoblastic leukemia. Haematologica. 2015; 100(6):786-793. https://doi.org/10.3324/haematol.2014.123273PubMedPubMed CentralGoogle Scholar
- Bassan R, Masciulli A, Intermesoli T. Final results of Northern Italy Leukemia Group (NILG) trial 10/07 combining pediatric- type therapy with minimal residual disease study and risk-oriented hematopoietic cell transplantation in adult acute lymphoblastic leukemia (ALL). Blood. 2016; 128(22):176. https://doi.org/10.1182/blood.V188.8.131.52PubMedGoogle Scholar
- Kantarjian HM, O'Brien S, Smith TL. Results of treatment with hyper-CVAD, a dose-intensive regimen, in adult acute lymphocytic leukemia. J Clin Oncol. 2000; 18(3):547-561. https://doi.org/10.1200/JCO.2000.18.3.547PubMedGoogle Scholar
- Kantarjian H, Thomas D, O'Brien S. Long-term follow-up results of hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone (Hyper- CVAD), a dose-intensive regimen, in adult acute lymphocytic leukemia. Cancer. 2004; 101(12):2788-2801. https://doi.org/10.1002/cncr.20668PubMedGoogle Scholar
- Del Principe MI, Buccisano F, Cefalo M. High sensitivity of flow cytometry improves detection of occult leptomeningeal disease in acute lymphoblastic leukemia and lymphoblastic lymphoma. Ann Hematol. 2014; 93(9):1509-1513. https://doi.org/10.1007/s00277-014-2080-6PubMedGoogle Scholar
- Del Principe MI, Buccisano F, Soddu S. Involvement of central nervous system in adult patients with acute myeloid leukemia: incidence and impact on outcome. Semin Hematol. 2018; 55(4):209-214. https://doi.org/10.1053/j.seminhematol.2018.02.006PubMedGoogle Scholar
- Mahmoud HH, Rivera GK, Hancock ML. Low leukocyte counts with blast cells in cerebrospinal fluid of children with newly diagnosed acute lymphoblastic leukemia. N Engl J Med. 1993; 329(5):314-319. https://doi.org/10.1056/NEJM199307293290504PubMedGoogle Scholar
- Roma A, Garcia A, Avagnina A, Rescia C, Elsner B.. Lymphoid and myeloid neoplasms involving cerebrospinal fluid: comparison of morphologic examination and imunophenotyping by flow cytometry. Diagn Cytopathol. 2002; 27(5):271-275. https://doi.org/10.1002/dc.10190PubMedGoogle Scholar
- Mitri Z, Siddiqui MT, El Rassi F. Sensitivity and specificity of cerebral fluid flow cytometry for the diagnosis of leukemic meningitis in acute lymphoblastic leukemia/lymphoma. Leuk Lymphoma. 2014; 55(7):1498-1500. https://doi.org/10.3109/10428194.2013.852667PubMedGoogle Scholar
- Wilson W, Bromberg J, Stetler-Stevenson M. Detection and outcome of occult leptomeningeal disease in diffuse large B-cell lymphoma and Burkitt lymphoma. Haematologica. 2014; 99(7):1228-1235. https://doi.org/10.3324/haematol.2013.101741PubMedPubMed CentralGoogle Scholar
- Nuckel H, Novotny JR, Noppeney R, Savidou I, Duhrsen U.. Detection of malignant haematopoietic cells in the cerebrospinal fluid by conventional cytology and flow cytometry. Clin Lab Haem. 2006; 28(1):22-29. https://doi.org/10.1111/j.1365-2257.2006.00741.xPubMedGoogle Scholar
- Sayed D, Badrawy H, Ali AM, Shaker S.. Immunophenotyping and immunoglobulin heavy chain gene rearrangement analysis in cerebrospinal fluid of pediatric patients with acute lymphoblastic leukemia. Leuk Res. 2009; 33(5):655-661. https://doi.org/10.1016/j.leukres.2008.09.033PubMedGoogle Scholar
- Cancela CSP, Murao M, Assumpção JG. Immunophenotyping of the cerebrospinal fluid as a prognostic factor at diagnosis of acute lymphoblastic leukemia in children and adolescents. Pediatr Hematol Oncol. 2017; 34(2):53-65. https://doi.org/10.1080/08880018.2017.1313920PubMedGoogle Scholar
- Ranta S, Nilsson F, Harila-Saari A. Detection of central nervous system involvement in childhood acute lymphoblastic leukemia by cytomorphology and flow cytometry of the cerebrospinal fluid. Pediatr Blood Cancer. 2015; 62(6):951-956. https://doi.org/10.1002/pbc.25363PubMedGoogle Scholar
- Martinez-Laperche C, Gomez-Garcia AM, Lassaletta A. Detection of occult cerebrospinal fluid involvement during maintenance therapy identifies a group of children with acute lymphoblastic leukemia at high risk for relapse. Am J Hematol. 2013; 88(5):360-365. https://doi.org/10.1002/ajh.23407PubMedGoogle Scholar
- Benevolo G, Stacchini A, Spina M. Final results of a multicenter trial addressing role of CSF flow cytometric analysis in NHL patients at high risk for CNS dissemination. Blood. 2012; 120(16):3222-3228. https://doi.org/10.1182/blood-2012-04-423095PubMedGoogle Scholar
- Subira D, Castanon S, Roman A. Flow cytometry and the study of central nervous disease in patients with acute leukaemia. Br J Haematol. 2001; 112(2):381-384. https://doi.org/10.1046/j.1365-2141.2001.02505.xPubMedGoogle Scholar
- SAlvarnas JC, Brown PA, Aoun P. Acute lymphoblastic leukemia, version 2.2015. J Natl Compr Canc Netw. 2015; 13(10):1240-1279. Google Scholar
- SArber DA, Borowitz MJ, Cessna M. Initial diagnostic workup of acute leukemia: guideline from the College of American Pathologists and the American Society of Hematology. Arch Pathol Lab Med. 2017; 141(10):1342-1393. https://doi.org/10.5858/arpa.2016-0504-CPPubMedGoogle Scholar
- SAkers SM, O'Leary HA, Minnear FL. VE-cadherin and PECAM-1 enhance ALL migration across brain microvascular endothelial cell monolayers. Exp Hematol. 2010; 38(9):733-743. https://doi.org/10.1016/j.exphem.2010.05.001PubMedPubMed CentralGoogle Scholar
- SAkers SM, Rellick SL, Fortney JE, Gibson LF. Cellular elements of the subarachnoid space promote ALL survival during chemotherapy. Leuk Res. 2011; 35(6):705-711. https://doi.org/10.1016/j.leukres.2010.12.031PubMedPubMed CentralGoogle Scholar
- Svan der Velden VH, de Launaij D, de Vries JF. New cellular markers at diagnosis are associated with isolated central nervous system relapse in paediatric B-cell precursor acute lymphoblastic leukaemia. Br J Haematol. 2016; 172(5):769-781. https://doi.org/10.1111/bjh.13887PubMedGoogle Scholar
- SFrishman-Levy L, Izraeli S.. Advances in understanding the pathogenesis of CNS acute lymphoblastic leukaemia and potential for therapy. Br J Haematol. 2017; 176(2):157-167. https://doi.org/10.1111/bjh.14411PubMedGoogle Scholar
- SPui CH, Howard SC. Current management and challenges of malignant disease in the CN in paediatric leukaemia. Lancet Oncol. 2008; 9(3):257-268. https://doi.org/10.1016/S1470-2045(08)70070-6Google Scholar
Figures & Tables
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.