1. Making sense of prognostic factors
Dr Dean Smith
City Hospital, Nottingham
22/03/17

2. Prognosis In Myeloma
Outcomes in myeloma have greatly improved over last two decades
Response to treatment and survival remains heterogeneous
A greater understanding of disease biology and mechanisms of resistance has led to an increased number of prognostic markers
With the introduction of novel therapies, the prognostic value of certain markers has varied according to which therapy the patient receives
Aim of this talk is to summarise current concepts of prognostic markers in myeloma

3. Categories of Prognostic Markers
Patient Characteristics
Disease Biology
Genetic Lesions
Gene Expression
Quality and Depth of Response
Radiological Features and effect of therapy
Staging Systems

4. Patient Characteristics

5. IMWG Frailty Score
Uses 3 geriatric assessment tools combined with age to categorise elderly patients into 3 groups: fit, intermediate fitness and frail
The assessments were performed on 869 newly diagnosed elderly patients in three prospective trials, using novel agents, predicting survival.
Long-term outcomes. (A) OS, (B) PFS, and (C) cumulative incidence of hematologic adverse events, (D) nonhematologic adverse events, and (E) discontinuation in the intention-to-treat population.
©2015 by American Society of Hematology
Antonio Palumbo et al. Blood 2015;125:

6. Disease Biology

7. Proliferation
Plasma cell labelling index – measures % cells in S-phase by fluorescent microscopy –traditionally showed that 1-3% of cells in S phase conferred a poor prognosis1
Recently, a high proliferation index (>1% in S phase) determined by flow cytometry has proven to be a prognostic tool
Paiva et al examined newly diagnosed transplant eligible patients in 2 clinical trials:
High Proliferation Index Predicts For Shorter OS
Bortezomib Based Induction May Abrogate This Effect
PFS and OS of patients with symptomatic MM submitted to HDT/ASCT grouped according to the presence of <1% and ≥1% S-phase PCs as assessed by MFC at diagnosis. A and B: PFS and OS of patients with MM included in the GEM2000 protocol (n = 319). C and D: PFS and OS of patients with MM included in the GEM2005<65y trial (n = 276).
A subgroup with proliferation index ≥ 3% had median OS 45 months
1 Griepp et al, 1996; Blood 81:
Paiva et al The American Journal of Pathology  , DOI: ( /j.ajpath )

8. Circulating Plasma Cells
Plasma cell leukaemia – primary and secondary - poor prognosis
Lack of prospective studies. 7 retrospective studies, pre novel agents, showed 5 year OS <10% .
Circulating plasma cells, not meeting diagnostic
criteria for PCL, still associated with poorer outcome in
a recent study
157 newly diagnosed patients treated novel agents
MFC used to detect circulating plasma cells
≥400 circulating plasma cells per 150,000 events
was associated with shorter OS
1 Fernandez de Larrea et al, 2013 Leukaemia; 27:
2 Gonsalves et al, 2014 Leukaemia; 28;

9. Genetic Lesions

10. Prognostic Cytogenetic Abnormalities In Newly Diagnosed Myeloma
STANDARD RISK
HIGH RISK
CYTOGENETIC ABNORMALITY
INCIDENCE (%)
t(11;14)1 15
t(4;14)2
t(6;14) 5
t(14;16)
2-3
HYPERDIPLOIDY 50
t(14;20) 1
Del(17p) 10
Del(1p)
Gain(1q)
35-40
Response rates often similar in patients with high risk and standard risk cytogenetics, but high risk progress earlier.
1t(11;14) – heterogeneity in Total Therapy Trials (TT2 and TT3). Patients with t(11;14) co-expressing CD20 had higher CR rate, quicker time to CR but shorter OS than those without CD20
2t(4;14) - poor risk can be (partly) overcome by bortezomib based therapy

11. Copy Number Abnormalities
Hyperdiploidy
Associated with standard outcomes
Retrospective Mayo Clinic study suggested trisomies abrogated poor prognosis conferred by adverse cytogenetic lesions1
These findings were not replicated in MRC Myeloma IX study2
Del(17p)
A strong poor prognostic factor in numerous different treatment regimens
HOVON/GMMG-HD4 trial suggested PAD→ tandem auto → bortezomib maintenance could partly overcome the poor prognosis of del(17p)3
Cavo et al ASH 2016: showed tandem autograft conferred a 20/12 improvement in PFS compared to single autograft for patients with high risk genetics4
STaMINA trial (ASH 2016) showed no benefit for tandem autograft or VRD consolidation compared to single autograft for high risk patients.5
1Kumar et al, 2012 Blood; 119:
2Pawlyn et al, 2015 Blood; 125:
3Neben et al, 2012 Blood; 119:
4Cavo et al, 2016 Blood; 128: 991
5Stadtmauer Late Breaking Abstract ASH 2016

12. Patients treated with bortezomib/dexamethasone then autograft:
Percentage of cells carrying del(17p) may be important – a 60% cut-off was suggested by the IFM1.
Patients treated with bortezomib/dexamethasone then autograft:
Subsequent trials have used lower cut-offs
The cut-off that is of significance has not yet been universally accepted, even by the most recent IMWG consensus paper on risk stratification2
Area that needs clarification going forwards for all cytogenetic
abnormalities
Poor prognostic impact of del(17p) only seen if present in over 60% of cells
1Avet-Loiseau et al, 2010 JCO; 28:
2Chng et al, 2014 Leukaemia; 28:

13. Chromosome 1 Copy Number Abnormalities
1p Deletion
1q Gain
Patients with >3 copies of 1q have a worse outcome, suggesting a possible dosage effect for the genes carried on 1q
Adverse impact of del(1p) is largely described in the context of ASCT
Myeloma IX
Intensive Treatment
Myeloma IX
Non-Intensive Treatment
HOVON/GMMG-HD4 trial
PAD vs VAD
Kai Neben et al. Blood 2012;119:
Pawlyn et al, 2016 Blood; 128: 1144
Kevin D. Boyd et al. Clin Cancer Res 2011;17:
Brian A. Walker et al. Blood 2010;116:e56-e65

14. Impact of Individual and Multiple Adverse Genetic Lesions Detected by FISH– Data From MRC Myeloma IX Trial
Prognostic Impact Of Adverse Genetic Lesions In Isolation
Prognostic Impact of Multiple Adverse Genetic Lesions
Adverse IgH translocation [t(4;14), t(14;16) and t(14;20)], del(17p) and 1q+ each had similar adverse impact on OS
Accumulation of adverse genetic lesions associated with shorter OS
Boyd et al, 2012 Leukaemia; 26:

15. Recurrent Mutations
Activating mutations are found in a number of genes in myeloma
NRAS 20-30% KRAS 15%
BRAF 5% CCND1 12%
Genes NKF-кB pathway 6-17%
NRAS and KRAS mutations conferred poor prognosis in studies pre-dating novel agents, but were prognostically neutral in Myeloma XI trial1
In Myeloma XI trial a number of mutations were prognostic1
CCND1 – associated with t(11;14). 2 year OS 38% mutated vs 80% non-mutated
Mutations preventing apoptotic signalling (i.e. ATM, ATR, TP53) conferred poor prognosis
Some mutations were associated with favourable OS, e.g. in IRF4 and EGR1
1Walker et al, 2015 JCO; 33:

16. NRAS and KRAS not associated with OS3
Mutational load and sub-clonal diversity has also been shown to be associated with poor outcome (Myeloma XI)1
MMRF CoMMpass trial: longitudinal study of 1147 newly diagnosed MM patients. Undertake genomic analysis at diagnosis and each progression event:
High mutational load, leading to high neo-antigen load, negatively impacts on survival2
Patients with one or two functional TP53 alleles had similar PFS and OS
Patients with zero functional TP53 alleles [i.e. del(17p) plus TP53 mutation, seen in c. 30% of del(17p)] had a significantly reduced OS (p<0.05)4
NRAS and KRAS not associated with OS3
1Melchor et al, 2014 Blood; 124: 640
2Miller et al, 2016 Blood 128; 193
3Keats et al, 2016 Blood; 128: 194

17. Gene Expression

18. Gene Expression Profiling
Several GEP models to risk stratify myeloma patients, e.g. Arkansas 70 and 17 gene high risk signatures1 and IFM-15 model2
EMC-92 GEP signature. Developed using HOVON-65/GMMG-HD4 data set (trial PAD vs VAD induction → auto → Thal vs BTZ maintenance)3
Validated in a variety of data sets, irrespective of age, treatment regimen and relapse status
In each case a HR population defined accounting for around 20% of the patient population
MRC-IX
transplant eligible
MRC- IX
transplant ineligible
APEX
Standard Risk OS
High Risk
1Shaughnessy et al, 2007 Blood; 109: Decaux et al, 2008 JCO; 26: Kuiper et al, Leukaemia; 26:

19. Integrated Genetic And Gene Expression Profiling
221 patients from transplant eligible arm of Myeloma XI were analysed for molecular genetic abnormalities and by GEP using EMC92 model
Patients with:
≥2 high risk genetic features [t(4;14), t(14;16), del(17p) or gain(1q)]
AND high risk by EMC-92 GEP signature
had significantly poorer outcome than patients with one of these features alone1
Combining genetic and gene expression profiling can identify a group of patients with “ultra high risk” disease
Median OS 25.6 vs vs. not reached;
p<0.0001
1Sherborne et al, 2016 Blood; 128:4407

20. Quality and Depth of Response

21. Quality and Depth of Response
A number of studies have established the link between deeper response and improved OS
IFM showed achieving at least a VGPR was associated with prolonged PFS in patients treated with bortezomib and dexamethasone or VAD.1
Further studies have shown achieving CR, rather than VGPR, is associated with longer OS2,3
Kapoor et al4 reported benefit of sCR over CR post autograft. Subsequent studies have suggested sCR, especially SFLC ratio, adds no additional prognostic information.5
344 patient post ASCT in Spain
No sig difference in OS between those achieving nCR, VGPR or PR
1Moreau et al, 2011 Blood; 117: Gay et al, 2011 Blood; 117: Martinez-Lopez 2011 Blood; 118:
4Kapoor et al, 2013 J Clin Oncol; 31: Martinez-Lopez et al, 2015 Blood; 126:

22. Early Relapse Post ASCT Predicts Poor Outcome
Shown in numerous studies
Single centre analysis of ASCT outcomes in 338 consecutive patients showed early relapse post ASCT (i.e. within 12 months) was the most important predictor of survival1
Patients with early relapse pre-2005 had similar outcome to those with early relapse post-2005
Suggests the adverse outcome of relapsing early was not salvaged using novel agent based treatment strategies at relapse
1Maciocia et al, Blood; 122: 2159

23. Minimal Residual Disease Predicts Survival
MRD Independently Predicts Survival per Log Reduction1
MRD Predictor Of Outcome In Transplant Ineligible Patients – Results From Myeloma XI Study2
397 patients in intensive arm Myeloma IX
Sequential improvement in OS for each log depletion in MRD level
14% of patients achieved MRD negativity
PFS benefit of 16 months [median PFS 34 vs 18/12 (p<0.001)]
Prognostic impact of MRD was independent of which treatment arm patients received
1Rawstron et al, 2015 Blood; 125:
2de Tute et al, 2016 Blood; 158: 245

24. Radiological Features and Effect of Therapy

25. Radiological Features and Effect Of Therapy
Imaging at diagnosis ≥3 focal lesions, the degree of uptake and extramedullary disease on PET-CT associated with shorter survival in an Italian study: Thal/Dex followed by ASCT. Imaging changes post treatment More rapid resolution of lesions is seen in PET-CT compared to MRI. In TT3, >3 Pet positive lesions on D7 of induction associated with poorer OS2 IFM/DFCI 2009 Trial (VRDx8 vs VRDx3/Auto/VRDx2) followed by lenalidomide maintenance. Showed PET negativity associated with improved PFS and OS3
OS according to PET negativity vs PET positivity before maintenance therapy
30-month OS rate: 69.9% in patients PET positive vs 94.6% in patients PET negative, p =0.01
1Zamagni et al, 2011 Blood; 118:
2Usmani et al, 2013 Blood; 121:
3Moreau et al, 2015 Blood; 126: 395

26. Staging Systems

27. REVISED- INTERNATIONAL STAGING SYSTEM (R-ISS)
INTERNATIONAL STAGING SYSTEM (ISS)
STAGE I
STAGE II
STAGE III
Serum β2 microglobulin <3.5 mg/l AND Albumin ≥35 g/l
Not fitting criteria for stage I or II
Serum β2 microglobulin ≥5.5 mg/l (irrespective of albumin)
Median OS 62 months
Median OS 44 months
Median OS 29 months
Retains prognostic value for conventional chemotherapy, high dose therapy and novel agents1,2,
Is prognostic for OS at relapse3
REVISED- INTERNATIONAL STAGING SYSTEM (R-ISS)
Combines the traditional ISS with presence or absence of high risk cytogenetics and normal or high serum LDH
Revised-International Staging System (R-ISS)4
ISS I + standard CA or iFISH or LDH normal
Not I or III
ISS III + high risk CA or iFISH or LDH high
High Risk FISH
17p del
and/or t(4;14)
and/or t(14;16)
With the advent of novel agents, IMIDS and PI, been increased OS in MM. But we know that despite, this, high risk cytogenetic abnormalities are associated with shorter overall survival.
1Kastritis et al, 2009 Leukaemia; 23: Munshi et al, 2011Blood; 117:
3Kumar et al, 2012 Leukaemia; 26: Palumbo et al. JCO doi: /JCO

28. R-ISS
Data was pooled from 4,445 patients with NDMM enrolled onto 11 international multicentre trials. 95% treated with novel agents
Using the R-ISS model:
38% stage 1, 62% stage 2, 10% stage 3
Prognostic for patients treated intensively and non-intensively
Limitations of R-ISS
chromosome 1 abnormalities not included – data not available
FISH analysis cut-off levels were not identical, ranged from 8% to 20% for del(17p) and from 10% to 15% for IgH translocations.
Does not include host-related prognostic factors such as age, performance status, and comorbidities

29. Summary of Prognostic Factors
BASELINE
POST-TREATMENT
TUMOUR BIOLOGY
Genetics1
Proliferative rate and circulating plasma cells
Extramedullary disease
(Further genetic changes)2
Depth of response
Timing and tempo of relapse
TUMOUR BURDEN
Beta-2 macroglobulin1
LDH1
(CR, MRD status)
PET-CT response
PATIENT FACTORS
Age
Albumin1
Frailty
Treatment emergent toxicity
1 Incorporated into R-ISS scoring system
2 Acquisition of additional genetic lesions at relapse, such as del(17p) or 1q amplification, is indicative of worse prognosis

30. How to make sense of prognostic information?
R-ISS is an attempt to integrate host and tumour related factors
Adverse features at baseline, e.g. extra-medullary disease may be used to “upstage” a patient into a higher risk category
Patients who relapse early clearly have high risk disease
Other markers, e.g. GEP and MRD analysis, are not routinely available at present, but need to be aware of these for trials and future use
With increasing choice of ever improving regimens, and increasing prognostic information, risk adapted strategies need to be assessed in prospective clinical trials
Such strategies should signify the first step towards individualised and more efficacious therapy