Analysis of long-term post auto-SCT outcomes

As the treatment landscape for multiple myeloma (MM) evolves, regular reviews are required to monitor the long-term effects on patient outcomes. Concomitant with improved treatment options and patient outcomes, the therapeutic aim has changed, with a greater emphasis on maximizing the depth of response to delay relapse and control the disease in the long term. Additionally, achieving MRD-negativity and sustained long-term disease control, referred to as “functional” or “operational” cures, may be more likely with the combination of novel agents and autologous stem cell transplant (auto-SCT). This is in contrast with the previous goal of “eradication” cure.

Auto-SCT is considered a crucial component of MM therapy and postulated to have a curative potential based on the graft-versus-myeloma effect.² To assess the improvements in long-term survival and rates of functional cures attained after a total therapy (TT) approach using induction therapy, auto-SCT, consolidation, and maintenance, in combination with novel agents, Katherine K. Nishimura, Cancer Research and Biostatistics, Seattle, US, and colleagues conducted a retrospective study.¹

Study design¹

• Patients (N = 4,329) with newly diagnosed MM who received auto-SCT as a frontline therapy at the University of Arkansas for Medical Sciences between 1989 and 2018 were included (Table 1). Patients were grouped into five time periods, that correspond to TT trials when novel therapies were first introduced, to determine if survival of patients who received auto-SCT had improved over time:
• < 1997 (reference group)
• 1998–2003
• 2004–2008
• 2009–2013
• ≥ 2014
• Patients were divided by molecular subgroups, determined by the translocation cyclin D (TC6) classifications
• Patient risk status was defined by the Gene Expressing Profiling 70 (GEP70) score

Table 1. Patient characteristics with significant differences by year of first ASCT¹

		Year of first auto-SCT
	All patients	< 1997	1998–2003	2004–2008	2009–2013	≥ 2014	p*
Auto-SCT, autologous stem cell transplant; GEP, Gene Expressing Profiling; ISS, International Staging System; TC6, translocation cyclin D; TT, total therapy; *p value connotates statistical significance
Sample size	4,329	661	1,002	1,294	837	535
Median follow-up time, y	10.5	21.5	15.3	11.0	6.5	2.5	—
Age, y							< 0.0001
Median (range)	58.9 (17.4–84.8)	53.0 (27.1–77.0)	57.4 (25.1–84.8)	59.7 (30.3–84.5)	61.7 (17.4–82.5)	63.0 (32.9–79.3)
< 65, %	72.1	90.6	77.2	70.3	63.7	57.6
≥ 65, %	27.9	9.4	22.9	29.7	36.3	42.4
On a TT clinical trial, %							< 0.0001
Yes	43.0	30.9	51.6	43.1	55.3	22.2
No	57.0	69.1	48.4	56.9	44.7	77.8
GEP70, %							< 0.0001
Low risk	42.5	0	23.0	54.6	70.6	58.1
High risk	9.2	0	3.3	13.1	14.8	13.3
No data	48.3	100	73.8	32.3	14.6	28.6
Any chromosomal abnormality, %							< 0.0001
Yes	37.9	68.2	62.9	64.5	52.6	53.8
No	61.1	28.6	36.0	34.9	47.1	46.0
No data	1.0	3.2	1.1	0.7	0.4	0.2
Tandem transplant, %							< 0.0001
No	38.1	37.4	39.1	25.9	46.1	53.8
Yes	61.9	62.6	60.9	74.1	53.9	46.2
Race/ethnicity, %
White	86.0	92.3	98.2	85.6	81.2	80.6	< 0.0001
African American	10.2	6.1	8.1	9.9	13.8	13.8
Other	3.8	1.7	2.7	4.5	4.8	5.6
TC6 classification							< 0.0001
CCND1	10.0	0.0	3.3	13.4	17.9	14.2
CCND3	1.0	0.0	0.5	0.9	2.3	1.3
D1	15.0	0.0	5.6	19.7	26.4	22.1
D2	14.0	0.0	4.8	19.0	22.2	22.2
MAF/MAFB	3.6	0.0	0.6	5.3	5.3	7.1
MMSET	6.3	0.0	2.9	8.7	11.0	7.7
No data	50.0	100	82.3	32.8	14.2	25.4
ISS							< 0.0001
I	42.6	43.9	54.1	40.4	32.3	39.0
II	33.6	28.3	26.1	37.6	40.6	34.1
III	18.3	11.0	18.6	20.3	24.4	12.2
No data	5.5	16.8	1.3	1.6	1.6	14.7

Results¹

With each successive time period and the integration of novel therapies, patients with MM that were treated with auto-SCT had an improved long-term survival. However, the elderly population (≥ 65 years old) and patients with high-risk disease still had a poor prognosis and survival.

• The median overall survival (OS) for all patients at a follow-up of 10.5 years (range, 0.01–26.4) was 6.9 years
• The 5-year Kaplan-Meier OS estimates show OS improved with each subsequent time period, p < 0.0001:
  • < 1997: 47%
  • 1998–2003: 58%
  • 2004–2008: 62%
  • 2009–2013: 61%
  • ≥ 2014: 70%
• Pairwise comparisons of OS between time periods did not show statistical significance, apart from the ≥ 2014 group who had superior OS vs all earlier time periods
• There was an improvement in the 5-year progression-free survival (PFS) with advancing time periods, p < 0.0001
  • < 1997: 29%
  • 1998–2003: 50%
  • 2004–2008: 61%
  • 2009–2013: 54%
  • ≥ 2014: 68%
• 6% of patients had not progressed after 3 years of follow-up
• The relative excess risk (RER) analysis shows MM-related death decreased as time progressed in comparison with the reference group, after SEER life tables were used to correct for differences in normal mortality:
  • 1998–2003: RER = 0.77 (95% CI, 0.68–88)
  • 2004–2008: RER = 0.66 (95% CI, 0.58–75)
  • 2009–2013: RER = 0.65 (95% CI, 0.55–76)
  • ≥ 2014: RER = 0.30 (95% CI, 0.22–41)
• Early mortality rate ranged from 36.4% in the reference group (< 1997) to 27.5% in the 2009–2013 group
  • Early mortality decreased in all time periods except for the ≥ 2014 group (32.3%) and was significantly different across the time periods (p = 0.002)
• Statistically significant differences in the proportion of patients with early mortality were observed in the following subgroups:
  • High-risk GEP70 vs low-risk GEP70: 61.4% vs5% (p < 0.0001)
  • Chromosomal abnormalities (yes vs no): 41.7% vs8% (p < 0.0001)
  • Age ≥ 65 years vs < 65 years: 37.2% vs6% (p < 0.0001)
  • Patients not treated on a TT protocol vs treated on a TT protocol: 37.9% vs 9% (p < 0.0001)
  • There were also significant differences by TC6 classification (p < 0.0001)
• Cox modeling, adjusted for age, gender, race/ethnicity, and clinical trial enrollment status, demonstrated OS and PFS had improved as time progressed. Patients receiving first transplant ≥ 2014 had the best survival compared to the reference group, HR 0.35 (95% CI, 0.27–45), p < 0.001 (Table 2)

Table 2. Cox proportional hazard model with covariates of statistical significance¹

CI, confidence interval; HR, hazard ratio; TT, total therapy
Covariate	HR	95% CI	p value
Year of transplant < 1997 1998–2003 2004–2008 2009–2013 ≥ 2014	Reference 0.08 0.69 0.68 0.35	— 0.72–0.89 0.62–0.77  0.59–0.78 0.27–0.45	— < 0.001  < 0.001  < 0.001  < 0.001
Age, years < 65 ≥ 65	(reference)  1.65	— 1.51–1.81	— < 0.001
Clinical trial Non-TT TT participant	(reference)  0.59	— 0.54–0.64	— < 0.001

• In patients with a complete response (CR), the percent of patients with CR or better at 5 years was
  • < 1997: 39%
  • 1998–2003: 60%
  • 2004–2008: 63%
  • 2009–2013: 60%
  • ≥ 2014: 32%
  • Patients treated after 1997 were more likely to have ≥ 5 years of CR duration and were less likely to relapse after achieving a CR
  • With longer follow-up, this may indicate patients are achieving a deeper response to treatment and preventing disease recurrence for longer
• Analysis of the effect of cytogenetics and tandem transplant:
  • Over time, no improvements in OS or PFS for single transplant vs tandem transplant recipients were observed
  • Patients with high-risk cytogenetics had a lower OS and PFS, but there were no differences between specific genetic abnormalities
• Curability:
  • Statistical cure fractions varied by patient subset
  • For the total cohort, the statistical cure fraction was estimated to be 16.4%
  • This increased over time from 6.3% pre-1997 to 31.3% for transplants 2009–2013
  • For many subgroups, statistical cure fractions were higher, with many patients achieving normal life expectancies with long-term event-free survival (EFS)
  • However, for patients with high-risk GEP70 and older patients, both the EFS and life expectancy were lower, this gives evidence a functional cure may exist

Conclusion

In summary, this study was able to cumulate substantial patient data and enabled comparisons of different treatment eras to assess the efficacy of auto-SCT in combination with novel agents used during induction, consolidation, and maintenance. The results demonstrate the combination of auto-SCT and novel agents were able to prolong OS and PFS whilst decreasing early mortality. It also highlights that high-risk patients did not significantly benefit in survival outcome. However, further periodic follow-up is warranted to be able to adequately test for improvements in longer-term survival, particularly in the > 2014 group that were assessed.

The introduction of thalidomide and bortezomib into standard clinical treatment significantly improved the outcomes of patients who receive auto-SCT. However, following an institute-wide shift from lenalidomide to thalidomide as standard therapy in the subsequent years, the authors were not able to replicate these findings and deduce a significant improvement. In more recent time periods, OS and PFS were superior for low-risk and older patients.

Although younger patients have better outcomes compared to older patients, both the Kaplan-Meier and Cox models show a promising improvement in outcomes of older patients in each consecutive year, with an increase in the 5-year survival rate from 35% to 63% and a reduction in mortality. This may be attributed to an earlier initiation of treatment based on improved detection or a more aggressive disease management approach.

The authors reported a high rate of early mortality and worse statistical cure fraction in patients with high-risk GEP70. Therefore, future research that focuses on both the short-term and long-term survival in patients with high-risk disease is warranted.