Modern Aspects in Lymphoma Diagnostics Prognosis and Prediction. Alexandar Tzankov

Transcription

1 Modern Aspects in Lymphoma Diagnostics Prognosis and Prediction Alexandar Tzankov

2 Introduction Prognostic and predictive markers Conclusions

3 Introduction Prognostic and predictive markers Conclusions

4 Quiz for clinicians In my daily routine I ask the diagnostic disciplines for prognostic/predictive markers in lymphomas 1. No, I am not aware of such markers 2. No, they do not influence my decision 3. Sometimes, after reading a paper 4. Yes 5. Only in relapsed or treatment-resistant cases

5 Quiz for diagnosticians In my daily routine I test lymphomas for prognostic/predictive markers 1. No, I am not aware of such markers 2. No, nobody is interested in such markers 3. Sometimes, after reading a paper 4. Yes 5. Only if required by the physician

6 55 y, male, diagnostic splenectomy CD20 TRAP

7 52 y, female, diagnostic splenectomy CD5 Cyclin D1

8 Exact diagnosis - the best prognostic and predictive factor Splenic marginal zone lymphoma Good prognosis after splenectomy 50% probability of not needing further therapy Chance to be HBV or HCV+ (needing therapy) Splenic mantle cell lymphoma Poorer prognosis Need of staging Need of adjuvant chemotherapy

9 Mitterlchener 2006

10 Introduction Prognostic and predictive markers Conclusions

11 Prognostic markers Help to stratify patients for treatment by identifying patients with different risks of outcome and thus may improve their chance of survival fate May have medical economical aspects

12 Prognostic markers - requirements Categorize patients Reproducibly measurable Readily available assessment technology

13 WHO categorization prognostic value Lymphoma entities Sir Thomas Hodgkin Virchow ARP Gall/Mallory Rappaport Lukes/Collins IWF updated Kiel REAL/WHO WHO

14 Prognosis of T-cell lymphomas Savage 2008

15 Predictive markers Indicate sensitivity or resistance to a specific therapy Allow better identification of patients, who will positively respond to the therapy Minimizing unnecessary suffering from side effects Reducing cost to both the patient and the wider health community

16 Predictive markers - requirements Indicate specific response to a certain drug Reproducibly measurable Readily available assessment technology

17 WHO categorization predictive value Lymphoma entities Sir Thomas Hodgkin Virchow ARP Gall/Mallory Rappaport Lukes/Collins IWF updated Kiel REAL/WHO WHO

18 Cladribine is that efficient only in HCL Saven 1998

19 MZL of MALT are curable with antibiotics

20 Interactions between therapy and prognostic markers ABVD Van Spronsen 1996 COPP

21 Interaction between therapy and prognostic markers with R no R Canioni 2008

22 Prognosis and prediction in lymphomas, beyond WHO DLBCL Hodgkin lymphoma FL and MCL T-cell lymphomas

23 Diffuse large B-cell lymphoma (DLBCL)

24 Therapy - the main prognosticator

25 Quiz In my daily routine I rely on DLBCL prognosis and treatment choice on 1. Gut feeling 2. IPI 4. Stage 5. Specific disease subtypes 6. All above

26 DLBCL heterogeneity

27 IPI challenges to the gold standard In R-CHOP treated cases IPI can not identify patients with <50% chance of survival (Sehn 2007) IPI 3-5 have similar outcomes under R-CHOP, but not IPI 0-2 (Ziepert 2010) In low IPI patients, tumor bulk is a risk factor (Pfreundschuh 2006) In relapsing patients time to relapse is of more significant prognostic value than IPI (Gisselbrecht 2010)

28 Does morphology further matter? DLBCL, NOS Centroblastic Immunoblastic Anaplastic Plasmablastic T-cell-/histiocyte-rich Lymphomatoid granulomatosis (Grade 3)

29 Immunoblastic vs centroblastic 1997 centroblastic immunoblastic Engelhard 1997

30 Immunoblastic vs centroblastic 2010 Ott 2010

31 Ott 2010

32 Plasmablastic DLBCL Tzankov 2004

33 T-cell-/histiocyte-rich LBCL

34 T-cell-/histiocyte-rich LBCL

35 THRBCL are not NPLHL

36 THRBCL are not NPLHL Biasoli 2009 & Bouabdallah 2003

37 Lymphomatoid granulomatosis

38 Predictive importance of the morphological variants of DLBCL Poorer response of immunoblastic DLBCL to R-CHOP Lacking Rituximab target in plasmablastic DLBCL A role for immune modulation (EBV specific CTC) in plasmablastic DLBCL and lymphomatoid granulomatosis?

39 Genetic heterogeneity of DLBCL Monti 2005

40 Genetic heterogeneity of DLBCL Alizadeh 2000

41 Event-free survival of DLBCL

42 ABC - a predictive factor Dunleavy 2009

43

44 Predictive importance of the gene expression profile variants Selective sensitivity to Bortezomib (ABC > GCB) Lenalidomide (ABC > GCB) Temsirolimus (GCB, PTEN loss of function) Enzastaurin (PKC ABC without NF- B mutations) Fostematinib (Syk, ABC dependent on BCR signaling) Hampered by unavailability of detection technology for mass analyses and poor reproducibility

45 Immunohistochemistry as a surrogate for gene expression profiling Meyer 2011

46 Reproducibility of semiquantitative immunohistochemical stainings of the Choi algorythm between observers and institutions

47 Basel Oxford Bcl-6 GCET1

48 Results Lawrie 2012

49 Gutierrez-Garcia 2011

50 Why are phenotypic markers less robust? Tissue preservation bias Subjective evaluation Poorly working antibodies Poor choice of markers Fail to consider +/- controls Poor choice of cut-offs Dependence on the applied therapy

51 Cut-off scores of prognostic phenotypic markers in DLBCL Tzankov 2008

52 ROC analysis as a powerful tool to choose proper prognostic cut-offs Tzankov 2010

53 ROC analysis as a powerful tool chose proper prognostic cut-offs Tzankov 2010

54 Prognostic markers with Rituximab Rituximab abolished the prognostic effect of Bcl-2 expression in DLBCL (Mounier 2003) Rituximab abolished the prognostic effect of Bcl-6 rearrangements/expression (Winter 2006) HIF1 correlates with better PFS in DLBCL with R-CHOP, but not CHOP, since HIF can regulate CD20 expression (Ahmed 2009)

55 BCL6 breaks and prognosis

56 DLBCL, topographically-defined entities Primary DLBCL of the CNS (of testis) Primary mediastinal B-cell lymphoma Primary cutaneous DLBCL, leg-type

57 Primary DLBCL of CNS, decreased efficacy of Rituximab Pharmacokinetic effects of the blood-brain barrier

58 Quiz When choosing Rituximab treatment I 1. Apply it to all B-cell lymphomas 2. Always ask the diagnostician about CD20 expression 3. Ask the diagnostician about CD20 expression only in relapses 4. Apply it especially in MZL 5. Perform a HBV screening

59 CD20 expression in B-cell lymphomas

60 -cell markers in MZL differentiation

61 -cell markers Genetic reporgramming e.g. c-myc abnormalities Hoeller 2008

62 ALK1+ DLBCL

63 -cell markers, -cell markers Genuine negativity for the respective marker EBV Tzankov 2004

64 -cell markers Oncogenic loss of B-cell identity Adams 2008

65 DLBCL of the testis CD20 Ki67 FOXP1

66 DLBCL of immune-privileged sites Booman 2008

67 DLBCL of immune-privileged sites More aggressive than most DLBCL Frequent relapses at other immuneprivileged sites

68 Rosenwald 2003 Primary mediastinal B-cell lymphoma

69 HL and PMBL closely related Clinically Mediastinal mass, young adults, F > M Morphologically Sclerosis, Reed-Sternberg-like cells Phenotypically CD30, no surface immunoglobulins decreased expression of OCT.2, BOB-1 Genotypically gains at 9p24 (JAK2 and PD1L locus) gains at 2p16 (REL locus) Meier 2009

70 CD20 CD30 CD15 Traverse-Glehen 2005

71 CD30 OCT-2 Traverse-Glehen 2005

72 Definition - B-cell lymphoma with overlapping features of HL/DLBCL clinical, morphological, phenotypic - Transitional phenotype, usually CD45+ Preserved B-cell signature (CD79a, BOB-1, OCT.2) Expression of CD30 and CD15 True grey zone cases Composite synchronous cases Sequential cases

73 Histopathological diagnosis most important predictive marker Hodgkin lymphoma Risk-adapted ABVD- or BEACOPP with or without irradiation Primary mediastinal B-cell lymphoma (R-)CHOP, (R-)MACOP-B, R-EPOCH Adjuvant radiotherapy does not offer survival benefit 10 CR/11 grey zone patients, 44% required radiation for CR OS and PFS at 4 years 86% and 57%, respectively Only 3 CR/12 HL patients treated with CHOP Dunleavy 2009 Miller 1983

74 Problematic issues Very problematic for treating physicians Seductive category (basket for difficult cases) Irreproducible category Kappa among 7 experts = 0.35

75 Hoeller 2010

76 Take home messages HL and PMBLC are closely related clinically, morphologically, pheno- and genotypically But Are treated in different ways Grey zones should be kept as few as possible Detailed phenotypic examination may help to classify most cases properly CD79a, BOB1, p63, CD45, CD23 and LMO2 (+ in PMBL) Cyclin E, MUM1, CD15 and LMP/EBER (+ in chl)

77 Some more grey zone issues Grey zones HL/DLBCL are not limited to the mediastinum and not limited to the differential of PMBL Classical HL/TCRBCL Classical HL/EBV+ DLBCL of elderly and EBV+ LPL NLPHL/TCRBCL

78 Lymphomas with special clinical context Paediatric type FL Paediatric type NMZL EBV+ T-cell LPD of childhood EBV+ DLBCL of the elderly

79 EBV+ DLBCL of the elderly EBNA 2 EBV+ >50J No immune deficiency or previous lymphoma Exclusion of LyG, PBL, PEL, HL, IM Poor prognosis 8% of DLBCL in Asia, 2.5% in Europe Hoeller 2010

80 Chronic inflammation associated DLBCL Pyothorax-associated DLBCL DLBCL associated with infected vascular prostheses DLBCL in arthritic joints can be diagnosed only integratively in >70% associated with EBV and CD20-negative

81 Predictive importance of lymphomas with special clinical context Modulation of immune defects is a treatment option EBV-antiviral therapy in latency type III (EBNA 2+) EBV-specific CTC in all EBV-associated cases Explant of prostheses in analogy to ALCL with breast implants

82 Quiz When treating/diagnosing DLBCL I 1. Always ask for the MYC status 2. Knowledge of the MYC status does not influence my therapy decision 3. Ask for the MYC status only in young patients 4. Perform a CNS prophylaxis in MYC rearranged cases 5. Analyze MYC only in high Ki-67 instances

83 DLBCL, C-MYC and Burkitt DLBCL 20 x more common than Burkitt in adults Burkitt 3 x more common than DLBCL in children C-MYC rearrangements in 5-9% of DLBCL unpredictable by morphology (Obermann 2009) poor prognosis with CHOP very poor if accompanied by additional aberrations

84

85 DLBCL vs Burkitt mbl index >.95 mbl index <.05

86 DLBCL vs Burkitt Hummel 2006

87 Open questions Is there still a prognostic role of the C-MYC gene status in R-CHOP treated DLBCL? Does the type of C-MYC gene aberrations play a distinct prognostic role?

88 Study design 601 patients with de novo DLBCL treated with R-CHOP and 332 patients treated with CHOP all with clinical follow-up and GEP data Interphase FISH for C-MYC gene abnormalities C-MYC break-apart probe (BAP) IgH/C-MYC double-fusion probe (DFP)

89 Results CHOP 217/332 evaluable (65%) R-CHOP 455/601 evaluable (76%)

90

91

92

93

94 Results

95 Results

96 Discussion kb ebl Exon 1 Exon 2 Exon 3 sbl HIV BL alternative

97 Conclusions C-MYC aberrations are detectable in 10% of DLBCL more commonly in GCB Alternative C-MYC rearrangements account for 3% C-MYC aberrations detected by BAP add independent prognostic information for individual DLBCL risk estimation in R-CHOP treated cases

98 Courtesy of Prof. Klapper

99 Predictive importance of MYC aberrations in DLBCL Lessons from children MYC translocations in molecular Burkitt and nonmolecular Burkitt are not of negative prognostic value if treated like Burkitt MYC rearranged DLBCL are often GCB and R-CHOP resistant DA-EPOCH-R or Burkitt-therapies in adults?

100

101 DLBCL, progress through differentiation? We are the best prognostic and predictive factors Right diagnoses Proper therapies Different DLBCL variants and subtypes associated With distinct outcomes Distinct genetical background Differentiated treatment modalities no Rituximab in CD20-negative cases PBL, ALK+ DLBCL, PEL Smart small drugs? Immunomodulation? EBV-specific therapy? - MYC rearranged cases?

102 Follicular lymphoma (FL)

103 FL, prognostic/predictive factors FLIPI sustained its role in the R era (Hiddemann 2005, Buske 2006) In stage I and II FLIPI does not identify patients at risk, but PET does 90% of PET+ progress within 2 years compared to 20% of PET- (Zinzani 2007) BM involvement - contraindication for radiolabeled anti-cd20 approaches

104 Dave 2004

105 High FOXP3 cell numbers better outcome Tzankov 2008

106 Prognostic importance of PD1+ T-cells Carreras 2009

107 Muenst 2010

108 Rituximab abolishes the prognostic significance of macrophages Canioni 2008

109 Relander 2010

110 Mantle cell lymphoma (MCL)

111 GEP risk groups are reflected by Ki-67

112 Hodgkin lymphoma (HL)

113 Quiz When treating/diagnosing Hodgkin lymphoma I 1. Not interested in subtyping since treatment is the same 2. Treat classical and NLPHL in a different manner 3. Treat classical and NLPHL in a different manner only in stage I & II 4. Interested in subtyping only for prognosis

114 Classical HL or nodular paragranuloma? Reed-Sternberg cell

115 Classical HL or nodular paragranuloma? Reed-Sternberg cell

116 Best predictor in stage I & II Classical HL chemotherapy (ABVD) +/- radiation Nodular paragranuloma chirurgical excision +/- Rituximab +/- radiation (+)/- chemotherapy Fong 2009

117 HL, prognostic/predictive factors Stage, B-symptoms IPS identifies only patients at very high risk (Hasenclever 1998) For low risk patients - low serum albumin and extranodal site of prognostic value (Franklin 2000) PET Post-therapy - prognostic with 90% specificity and sensitivity (Zijlstra 2006) Interim - reliable to identify poor responders

118 HL, tumor-related prognostic factors - it is all in the background! Eosinophils better prognosis (Wassliewski 2000) Mast cells - worse prognossi (Canioni 2009) Macrophages worse prognosis (Steidl 2010) Granzyme B - worse prognosis (Oudejans 1997) FOXP3 + - better prognosis (Tzankov 2008) B-cells - better prognosis (Chetaille 2009)

119 Steidl 2011

120 A possible prognostic role of macrophages

121 Steidl 2010

122 CD68 low CD68 high Cut off: >1% with respect to OS prognostic sensitivity: 75% prognostic specificity: 65% Tzankov 2010

123 Prognostic significance of CD68- positive tumor macrophages in HL Tzankov 2010

124 FOXP3 Tzankov 2008

125 Prognostic significance of PD1-positive tumor-infiltrating lymphocytes in HL Muenst 2009

126 Prognosis of chl with respect to the cellular background composition Muenst 2009

127 Predictive importance for differential immunomodulation? Steroids, Rapamycin, allo-sctx

128 Do we need tumor-cell related prognostic/predictive factors in HL? Fong 2009

129

130 Problems with the specificity of HDI due to omnipresence of targets Adams 2010

131 Survival of HL after relapse

132 Every fifth relapse is clonally unrelated Obermann 2011

133 T-cell lymphomas (TCL)

134 Prognosis of T-cell lymphomas Savage 2008

135 TCL IPI is valid in T-NHL Prognostic index for PTLC (PIT) bone marrow involvement, age, ECOG and LDH (Gallamini 2004) Prognostic index for NK cell lymphomas non-nasal location, stage and ECOG (Suzuki 2010)

136 TCL, tumor related prognostic markers Ki-67 >80%, EBV and CD15: worse prognosis in PTCL (Went 2006) >70% blasts, Ki-67 >25%, CD56, CD30, EBV: worse prognosis (Rudiger 2002) Cytotoxic phenotype in PTCL worse prognosis (Asano 2009)

137 PTCL, adverse role of cytotoxic phenotype Iqbal 2010

138 Predictive markers in TCL ALK1+: from molecular subtype to Crizotinib PTCL high PDGFR Imatinib (Piccaluga 2007) AITL high VEGF Bevacizumab (Piccaluga 2007, Bruns 2005) CD30-negative TCL express CD52 Alemtuzumab (Enblad 2004) CD30 a therapeutic target for Iratumumab, SGN-30 & 35 (Ansell 2008, Younes 2010) R-CHOP in AITL with lots of EBV-infected B-cells? (Skugor 2010) Nasal NK lymphomas low asparagin synthetase L-asplaraginase (Jaccard 2011)

139 PDGFR and VEGF in PTCL and AITL PDGFR VEGF VEGFR2 Piccaluga 2007

140 Introduction Prognostic and predictive markers Conclusions

141 Prognosis and prediction in lymphomas Precise diagnosis - best prognosticator and predictor Lots of potential tumor- and micromilieu-related prognostic markers Rational molecular targets identified prospective targeted molecular trials

142 Quiz The presentation 1. Was informative but would not change my management strategy 2. Was bad 3. Would somehow influence my management strategy 4. Forces me to look at the recent literature

143

144 Entity Factor Importance ALCL Expression of ALK1 Better prognosis, response to Crizotinib B-NHL CD20 Target of monoclonal antibodies like Rituximab DLBCL FL HL ABC subtype MYC-gene rearrangement Increased regulatory and follicular T-cells Increased regulatory T-cells Increased cytotoxic T-cells Increased macrophages MCL High proliferation Worse prognosis Worse prognosis, response to smart drugs Worse prognosis, poor response to R- CHOP, potential for Burkitt therapy protocols Better prognosis, better response to Rituximab Better prognosis Worse prognosis Worse prognosis, poorer response to HD- ASCT when relapsed MZL t(11;18) Resistance to Helicobacter pylori eradication (Liu, 2001)