Utility or futility? A contemporary approach to allogeneic hematopoietic cell transplantation for TP53-mutated MDS/AML

TP53-mutated myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) are among the most lethal malignancies, characterized by dismal outcomes with currently available therapies. Approximately 6% to 10% of de novo MDS/AML cases harbor TP53 abnormalities,^1-6 and the genetic lesion is further enriched in therapy-related disease, affecting at least 20% to 35% of such cases.^1,7,8 The TP53 gene encodes for a tumor-suppressor protein that orchestrates transcriptional response to genotoxic stress, regulating cell cycling and DNA repair or apoptosis when damage is irreparable.⁹ Inactivating mutations in TP53, therefore, confer high resistance to conventional cytotoxic chemotherapy. More than half of the number of patients with TP53-mutated AML do not achieve remission, and the mutation is linked with short response duration in MDS treated with hypomethylating agents (HMAs).^10-14 Given their uniformly aggressive behavior, TP53-mutated myeloid neoplasms are now categorized as a biologically distinct entity by international expert panels.^15-17 Recent estimates demonstrate a 2-year survival of <15% among all patients diagnosed with TP53-mutated MDS/AML that has not improved in recent years even with the incorporation of newer therapies, including the oral Bcl-2 inhibitor venetoclax.^11,18 

Allogeneic hematopoietic cell transplantation (allo-HCT) is widely thought to be the only treatment option to offer durable disease control. However, outcomes with allo-HCT in this context are quite poor, calling into question the utility of transplantation in this biological subgroup. Only a minority of patients with TP53-mutated MDS/AML undergo allo-HCT.^11,12,18 Given a low probability of long-term success and the highly morbid and resource-intensive nature of transplantation, it is understandable that physicians may hesitate to recommend this treatment to their patients. In addition, transplantation centers are expected to meet certain benchmarks of patient survival and may be reluctant to perform such high-risk transplants given historically poor results.¹⁹ 

In this review, we contend with some of the most challenging clinical questions facing leukemia and transplantation physicians, that is, whether, when, and how to perform transplantation on patients with TP53-mutated MDS/AML. We summarize the latest data on allo-HCT outcomes in this subgroup, evaluating the limitations of available evidence; we review the molecular heterogeneity of this disease, delineating outcomes based on distinct biological features to aid in patient selection; and we critically examine whether allo-HCT should be routinely applied in this disease on the basis of currently available data.

Though prognoses of TP53-mutated MDS/AML are generally poor, biological heterogeneity confers some variation in treatment outcomes. Cytogenetic and molecular features may help in stratifying outcomes in the transplant setting.

TP53-mutated MDS/AML is often accompanied by adverse-risk cytogenetics, including loss of chromosome 17 and/or complex or monosomal karyotypes,²⁰ the combination of which is seldom cured with allo-HCT (Table 1). A registry-based analysis from the European Society for Blood and Marrow Transplantation evaluated outcomes after allo-HCT in 179 patients with TP53-mutated AML in first complete remission (CR).²¹ Strikingly, patients with mutant TP53 with no evidence of 17p loss or complex karyotype (n = 53; 26.9%) had a 2-year overall survival (OS) that was comparable with that of patients with preserved TP53 function (65.2% vs 64%; P value is not provided). Patients with mutant TP53 accompanied by chromosome 17p loss and/or complex karyotype (n = 126; 70.4%), however, experienced a 2-year leukemia-free survival of only 15% (P = .001). Relapses continued beyond the 2-year mark, with no detectable plateau on the survival curve for patients with the combination of mutant TP53 with 17p loss and/or complex karyotype. The presence of other cytogenetic abnormalities associated with adverse risk in AML including monosomal karyotype, deletion 5q or deletion 7q alongside mutant TP53 similarly predicted for poor survival after transplantation. Middeke et al²² and Grob et al¹⁸ both reviewed outcomes of patients with TP53 mutations treated with allo-HCT within several clinical trials of AML and AML/MDS-excess blasts, respectively, and each reported a 3-year OS of ∼10% in patients with concurrent adverse-risk cytogenetics. Older studies lacking molecular data but capturing the loss of TP53 function as a result of chromosome 17p abnormalities based on conventional cytogenetics and/or fluorescence in situ hybridization analysis similarly illustrate the prognostic relevance of coexisting cytogenetic abnormalities in predicting posttransplantation outcomes in AML.^23,24 

Results appear equally discouraging in patients with MDS. A Japanese study of 797 patients with MDS who underwent unrelated donor allo-HCT through the Japan Marrow and Donor Program found that among the 58 patients with TP53 mutations and complex karyotype, the median OS after HCT was only 4.3 months, with ∼60% of patients dying of early relapse.²⁶ Notably, the majority of patients with TP53-mutated MDS/AML have these additional adverse disease characteristics.^1,11,29 In the study by Grob et al, for example, 83% and 78% of patients with TP53 mutated MDS/AML had complex and monosomal karyotypes, respectively.¹⁸ Taken together, these data indicate that only a small minority of patients with TP53-mutated MDS/AML have a feasible chance at achieving long-term survival with current therapies that include allo-HCT.

Variant allele frequency (VAF) and allelic state also enables refinement of prognostic significance of TP53 alterations in the transplant setting. Biallelic loss of TP53 function occurs in the high majority of TP53-mutated MDS/AML^1,18,30 and can occur through a cytogenetic deletion of the 17p locus alongside a mutation in TP53, as reported earlier, ≥2 distinct TP53 mutations involving both alleles, or 1 mutation with VAF > 50%, indicating a loss of heterozygosity or hemizygosity.¹⁵ Short et al investigated the prognostic significance of VAF in 202 patients newly diagnosed with TP53-mutated AML.¹² In the subset of 20 patients who underwent allo-HCT, those with TP53-mutated VAF ≤ 40% at diagnosis had a trend toward better OS compared with patients with VAF > 40% (P = .07). Furthermore, patients with VAF > 40% who underwent allo-HCT did not achieve meaningfully improved survival compared with patients who did not receive an allograft (median OS, 9.8 and 8.0 months, respectively; P = .09). The survival benefit of allo-HCT was much more pronounced in patients with VAF < 40% at diagnosis (median OS, 32.3 and 9.5 months, respectively; P = .01). Not surprisingly, higher VAF was strongly associated with abnormal cytogenetics. In a single-institution study from Canada, investigators reported that patients who underwent transplant with lower mutant TP53 VAF (defined as <45%) at initial diagnosis had superior outcomes compared with those with higher TP53 VAF (hazard ratio [HR], 0.19; P = .02).³¹ Only patients with lower TP53 VAF benefited from allo-HCT when survival was analyzed in a landmark analysis to account for immortal time bias. Byrne et al retrospectively studied post-HCT outcomes of 384 patients with TP53-mutated MDS/AML; 74% of patients had biallelic targeting of the TP53 gene, and monoallelic mutations were associated with improved OS (HR, 0.52; P = .001).³² Notably, monoallelic missense TP53 mutations have been identified in clonal hematopoiesis. As such, somatic TP53 mutations particularly when present at low VAFs may represent pre-existing clonal hematopoiesis rather than leukemic subclones.^18,33 

Versluis et al recently performed genomic testing on enrollment samples of patients treated on Blood and Marrow Transplant Clinical Trials Network (BMT-CTN) 1102, a biological assignment trial of allo-HCT based on donor availability in older adults with high-risk MDS.^34,35 Investigators found that the survival benefit of allo-HCT observed in the trial extended to patients with TP53-mutated disease. A multivariable model applied to 80 patients with a TP53 mutation showed that patients in the donor arm had a nonsignificantly improved OS compared with those in the no donor arm (HR, 1.76; P = .073). Although outcomes were not stratified based on cytogenetic or molecular features, multihit TP53 mutations were shown to be associated with nonsignificant inferior survival compared with single-hit mutations (HR, 1.64; 95% confidence interval, 0.92-2.94; P value is not provided), when the model was adjusted for donor arm.

TP53 clearance is not easily achieved in MDS/AML with currently available induction therapies, and the depth of remission necessary to achieve long-term survival after transplantation is currently unknown. Importantly, detectable mutations do not necessarily represent measurable residual disease (MRD), because TP53 mutations may relate to pre-existing clonal hematopoiesis as noted above, making it difficult to contextualize available data.

In most patients, TP53 mutations remain detectable in complete morphologic remission at the time of allo-HCT. Grob et al performed deep-targeted sequencing on bone marrow samples from 62 patients with TP53-mutated MDS/AML in complete morphologic remission, and found that 45 patients still had detectable TP53 mutations (level undefined).¹⁸ Hunter et al found that fewer than half of the patients with TP53-mutated MDS who received treatment at their institution achieved TP53 mutation clearance to a VAF < 5%.¹³ Patel et al reported that only 1 in 40 patients at their institution with TP53-mutated MDS/AML achieved mutation clearance (level undefined) with induction therapy.¹⁹ 

Several small analyses have investigated whether the persistence of these mutations at the time of HCT contribute to poor outcomes in patients with MDS (Table 2). In the aforementioned Japan Marrow and Donor Program study of patients who underwent transplant for MDS, Yoshizato et al found that the fraction of TP53-mutated cells at the time of HCT significantly correlated with shorter time from transplantation to death.²⁶ Hunter et al performed serial molecular profiling on 47 patients with TP53-mutated MDS, treated with HMA-based therapies at their institution; 16 patients proceeded onto allo-HCT.¹³ Patients who underwent allo-HCT with TP53 persistence (defined as VAF ≥ 5%) suffered the same poor outcomes as patients who did not undergo allo-HCT and instead remained on HMA therapy. Patients who underwent allo-HCT with VAF < 5% experienced a survival benefit compared with patients who remained on HMA therapy (25.2 vs 7.7 months; P = .005). This prompted the authors to propose a treatment algorithm in which allo-HCT is recommended according to whether molecular remission is achieved in patients with TP53 mutated MDS. Patients with clearance of TP53 demonstrated lower median VAF at diagnosis (12% vs 33.3%), which might be expected to reflect a less aggressive disease phenotype, but no difference in frequency of biallelic inactivation or complex cytogenetics was seen. This may reflect the small numbers in this analysis. In contrast to these 2 studies, Versluis et al found that TP53 clearance before HCT (analyzed at thresholds of <5% and <2%) did not associate with posttransplant survival in older adults with MDS treated with reduced intensity conditioning (RIC) allo-HCT on the BMT-CTN 1102 biologic assignment trial.³⁵ 

Data regarding the prognostic significance of TP53 mutation persistence at time of allo-HCT are similarly conflicting in AML. In general, presence of MRD before allo-HCT is a well-established risk factor for disease relapse and inferior outcomes in patients with AML.^39-41 In the recent European Society for Blood and Marrow Transplantation registry analysis by Loke et al, the investigators found that detection of MRD (through whichever methodology that was performed by contributing centers) before allo-HCT resulted in inferior OS in patients with TP53-mutated AML (18-month OS, 64.5% vs 37.3%; P = .008).²¹ In contrast, data from the Consortium on Myeloid Malignancies and Neoplastic Diseases study demonstrated that MRD status before allo-HCT, as assessed by flow cytometry or by TP53 mutation clearance, did not significantly affect event-free survival.³⁸ 

In summary, the impact of TP53 mutational persistence on long-term posttransplant outcomes remains unknown. Interpretation of available data is hampered by small numbers of patients and variability in depth of sequencing and VAF cutoffs used, precluding any definitive conclusions regarding this question. Larger, prospective studies are needed to better define the depth of remission necessary to achieve long-term survival after transplantation in this disease subset, as well as the influence of varying intensity conditioning regimens on residual mutations.

Among TP53 mutational variants in MDS/AML, missense mutations clustering in the DNA-binding domain predominate.^42-44 Whether clinical outcomes differ between patients with missense mutations vs truncating aberrations, which include nonsense mutations, frameshift deletions or insertions, and splice site variants,⁴⁵ is not well understood. Lindsley et al examined the outcomes of 289 patients with TP53-mutated MDS who underwent allo-HCT and found that patients with only truncating TP53 mutations experienced shorter OS compared with patients with missense variants.²⁵ However, this has not been consistently demonstrated in the broader MDS/AML setting of variably treated patients^1,44,46 and, to our knowledge, little data otherwise exist on the functional classification of TP53 mutations in the setting of transplantation.

Not all analyses report such discouraging findings of transplantation outcomes in TP53-mutated MDS/AML, with some studies describing durable remission in a sizable subset of patients. Shahzad et al performed a meta-analysis of 6 studies including 248 patients with TP53-mutated AML who had undergone allo-HCT and reported a pooled 2-year OS of 30%.⁴⁷ Badar et al recently reported a study on 370 patients with TP53-mutated AML treated in the last decade via the Consortium on Myeloid Malignancies and Neoplastic Diseases. Among the 68 patients who received an allograft, the 4-year event-free survival was ∼25% in patients who underwent transplant in first remission.²² In the study by Lindsley et al, patients with TP53-mutated MDS who underwent allo-HCT achieved a 4-year OS of ∼20%.²⁵ Importantly, most of these studies did not stratify survival outcomes based on cytogenetics or allelic state, raising the question of whether those who achieved durable remission were primarily among the minority subset of patients with monoallelic TP53 mutations and without concurrent high-risk cytogenetics.¹ In addition, worth noting are 2 studies that did not detect any negative prognostic impact of TP53 mutations in the transplant setting. In a single-center analysis of 60 patients with therapy-related MDS, Aldoss et al found that the presence of a TP53 mutation, seen in 18 patients, was not detrimental to posttransplantation outcomes.⁴⁸ Similarly, Welch et al performed an institutional trial of 10-day decitabine induction regimen in 84 patients with MDS/AML and concluded that outcomes were not adversely affected by presence of TP53 mutation (n = 21). Posttransplant survival also did not differ according to the presence of a TP53 mutation. Notably, survival was short among patients enrolled on study regardless of mutational status (median OS, 12.7 months among patients with TP53 mutations and 15.4 months among patients with wild-type TP53; P = .79).⁴⁹ Findings from these 2 studies stand in sharp contrast to our own center’s experience, several registry analyses and large studies reporting the adverse prognostic impact of TP53 mutations on posttransplantation outcomes in MDS/AML.^{13,18,21,22,25,26,29,50,51} The reasons for these notable outcome discrepancies are unclear.

Patient health characteristics, including performance status and comorbidities, affect the risk of nonrelapse mortality (NRM) and influence survival after allo-HCT. The decision to proceed to allo-HCT in any setting is ultimately a balance of risk and benefit. When the likelihood of benefit is small or uncertain based on high rates of relapse, the consideration and quantification of risk based on physical health and comorbidities assume even greater significance. Ciurea et al reviewed outcomes of 83 patients with TP53-mutated MDS/AML who underwent allo-HCT at 1 institution and found 3 factors to be predictive of OS: HCT-comorbidity index (HCT-CI)⁵² > 4 (HR, 3.9; P = .03), Karnofsky performance status (KPS) ≤ 80% (HR, 3.04; P = .07), and more advanced disease not in first or second CR (HR, 4.1; P = .004).⁵³ Patients without any of these 3 risk factors achieved a 1-year progression-free survival of 40%, whereas only 8% of patients with ≥2 risk factors survived 1 year without relapse. HCT-CI and KPS predicted for increased risk of NRM, as did therapy-related disease (HR, 4.2; P = .007). Correspondingly, history of treated malignancy generates 3 points on the HCT-CI and strongly predisposes to NRM,⁵⁴ perhaps partly a reflection of physiological stress incurred by prior therapy. This is highly relevant in a biological subset of disease that is frequently therapy-related, so this point must be carefully considered when evaluating these patients for allo-HCT.³² Intriguingly, a recent study found TP53 mutations in MDS/AML to be associated with increased pretransplant infections including bacterial pneumonia and invasive fungal infection culminating in a higher rate of death from infection compared with patients with wild-type TP53 despite similar durations of neutropenia⁵⁵; whether TP53 alterations impart some biological predisposition to infection or phagocytic dysfunction is unknown. Large analyses on post-HCT outcomes in TP53-mutated MDS/AML do not show increased NRM compared with wild-type TP53, which may reflect stringent patient selection.^21,22,26 

Collectively, studies reporting on the allo-HCT experience in TP53-mutated MDS/AML illustrate a salient point, that is, only a minority of patients with this disease subtype undergo allo-HCT. In the aforementioned study by Short et al, allo-HCT was only undertaken in 21% of patients who achieved CR.¹² In the aforementioned study by Grob et al, only 26% of the 230 patients with TP53 mutant MDS/AML received an allograft.¹⁸ The reasons for not proceeding to allo-HCT in these studies are not described. In the analysis by Badar et al, only 18% of the 370 patients with TP53 mutated AML received an allograft. The authors reported on barriers to allo-HCT for those who did not undergo transplantation: 47% of patients did not achieve a response to induction and/or salvage therapy and, thus, did not proceed on the basis of inadequate disease control. Among responders (defined as achieving CR, CR with incomplete count recovery, partial remission, or morphologic leukemia-free state), 71% of patients did not proceed because of advanced age, comorbidities, or lack of donor availability. This highlights that the patients who do undergo allo-HCT are highly selected on the basis of health and/or fitness, factors which could account, at least in part, for the perceived survival benefit of this therapy. It is, therefore, our view that achievement of any favorable posttransplant outcomes in this select subgroup cannot be used to advocate for widespread use of allo-HCT in the broader population of patients with TP53 mutated MDS/AML, considering most of these patients do not exhibit optimal health characteristics.

TP53 mutations confer refractoriness to conventional cytotoxic chemotherapy in AML^56,57 and are associated with short response duration in MDS treated with HMA.^10-12 No novel agent or treatment regimen has improved long-term outcomes of TP53-mutated MDS/AML in recent years, including regimens containing venetoclax, the novel TP53-targeting agent eprenetapopt or the anticluster of differentiation 47 antibody magrolimab.^58-61 Comparative studies have yielded few results to guide the optimal pretransplant therapeutic strategy in patients with TP53-mutated MDS/AML. More broadly, the question of the optimal induction regimen in patients with AML planned to undergo allo-HCT remains unanswered.^62,63 One study suggested improved outcomes with venetoclax-based lower intensity regimens in patients with adverse-risk AML including TP53 mutations, but very small numbers and retrospective comparison preclude any major conclusions.⁶³ This finding warrants validation in larger multicenter cohorts.

To date, there has been no proven advantage of any specific transplant approach in this subset of disease with regard to selection of graft source, conditioning, T-cell manipulation, graft-versus-host-disease prophylaxis, or posttransplant maintenance. Ciurea et al found that conditioning regimen, donor type, and graft source were not predictive of outcomes.⁵³ Similarly, Loke et al reported that none of these transplant-related variables influenced outcomes.^21,TP53 mutations are known to confer chemoresistance, which may be relevant when selecting conditioning intensity.^57,64 However, Lindsley et al reported that myeloablative conditioning offered no advantage over RIC for patients with TP53-mutated MDS.²⁵ Byrne et al performed a multicenter analysis of 384 patients with TP53-mutated MDS/AML who received allo-HCT. Among 172 patients with MDS, RIC increased the risk of post-HCT relapse (HR, 2.54; P = .009) but OS was not affected. Conditioning intensity did not influence outcomes in 212 patients with TP53-mutated AML.³² In a single-center study, Chan et al found that in patients with MDS/AML with TP53 persistence at the time of allo-HCT, myeloablative conditioning led to inferior OS compared with RIC (HR, 2.6; P = .023).⁶⁵ Results of the randomized BMT-CTN 0910 support myeloablative conditioning in AML,⁶⁶ especially in the presence of MRD,⁴⁰ but analyses have shown no influence of conditioning regimen on cumulative incidence of relapse or survival in TP53-mutated AML specifically.^21,23,38 

Chronic graft-versus-host-disease development is associated with reduced risk of post-HCT relapse in TP53-mutated MDS/AML, supporting the susceptibility of this disease subset to a graft-versus-leukemia effect.^32,38 Nonetheless, available data have not demonstrated survival differences based on donor age or degree of HLA-disparity.^22-24 Loke et al found that matched unrelated donors, who are expected to be younger, may impart a relapse benefit over matched sibling donors (HR, 0.76; P = .12), but no OS difference was detected.²¹ A small, retrospective analysis suggested reduced relapse after cord-based vs matched donor transplants (1-year cumulative incidence of relapse, 31% vs 55%; P = .122), but the impact was nullified by increased NRM.⁶⁷ 

Finally, post-HCT maintenance strategies are appealing in a disease subset characterized by high rates of early relapse, but their benefit in this setting is highly speculative. More broadly, few posttransplant maintenance therapies have demonstrated benefit in randomized studies, and no novel agent or treatment regimen has improved outcomes of TP53-mutated MDS/AML.¹¹ Eprenetapopt has been studied in combination with azacitidine for post-HCT maintenance. It was deemed to be well-tolerated,⁶⁸ but the approach has not been tested in an randomized-controlled trial (RCT). There are several currently ongoing investigations of maintenance strategies post-HCT that may be relevant to this disease subset including use of oral HMAs (NCT04173533, NCT04980404) and the oral Bcl-2 inhibitor venetoclax (NCT03613532), as well as the phase 3 VIALE-T study of azacitidine combined with venetoclax in AML (NCT04161885). Though one might apply maintenance therapies in patients with TP53-mutated MDS/AML in an attempt to improve upon historically dismal outcomes, this approach should not be used to motivate or justify allo-HCT in this population outside of a clinical trial. A transparent discussion with patients regarding potential risks and unknown benefit would be required.

Given the absence of dedicated RCTs investigating this question, we believe the survival benefit of allo-HCT in TP53-mutated MDS/AML with biallelic loss and/or adverse-risk cytogenetics has not been established. Observational or retrospective studies must be interpreted with caution. The fittest patients with the most responsive disease are selected to undergo allo-HCT, and in order to do so, a patient must survive long enough from diagnosis to enter remission and receive an allograft. Judicious patient selection, as described earlier, confounds interpretation of the impact of allo-HCT. We believe allo-HCT may afford extended survival to or potentially even cure highly selected patients with TP53-mutated MDS/AML, but what is far less clear is how to prospectively select the minority of patients whom it might benefit. Adverse biological features such as TP53 mutations and adverse-risk cytogenetics are more prevalent in older adults, and advanced age itself may diminish any potential survival benefit associated with allo-HCT.⁶⁹ 

We believe that the exceptionally poor outcomes of patients with TP53-mutated MDS/AML with biallelic loss and/or adverse-risk cytogenetics should motivate RCTs of HCT vs non-HCT to determine whether transplantation can prolong survival and/or positively affect other clinically relevant outcomes, such as patient-reported outcomes or healthcare resource utilization in this disease subset. Arguments against performing such an RCT is the absence of clinical equipoise based on perceived benefit of allo-HCT. To our knowledge, the only randomized trial data in TP53-mutated AML strongly supports clinical equipoise: in an analysis of >3500 patients intensively treated in 3 RCTs of postremission therapy, 143 patients (4%) possessed chromosome 17p abnormalities. Forty-seven of these patients were consolidated with allo-HCT. Investigators found that allo-HCT did not confer a survival advantage over chemotherapy.⁵⁴ 

Our field must contend with any approach broadly offering an inherently morbid and resource-intensive procedure for a population when the overwhelming majority of these patients are expected to relapse or die early after transplant. Without dedicated prospective randomized trials, selecting who may actually derive benefit from allo-HCT for TP53-mutated MDS/AML can be described as ambiguous guesswork and must be carefully contemplated. Our current practice is to offer allo-HCT to patients with TP53-mutated MDS/AML who are least likely to suffer transplantation-related morbidity and mortality. We selected these patients on the basis of very good performance status (KPS ≤ 90) and minimal comorbidities^52,70,71 (Table 3). For patients not meeting these strict characteristics, we are unlikely to offer allo-HCT to patients with biallelic alterations and/or adverse-risk cytogenetics. For all patients with TP53-mutated MDS/AML, we are extremely clear about high rates of early relapse and likely need for posttransplantation therapy.

Until more effective agents to treat TP53-mutated myeloid malignancies are developed, perhaps the most impactful thing oncologists can do when considering allo-HCT for patients afflicted with this disease is to set realistic expectations. We believe this requires transparency with patients regarding outcomes of allo-HCT, and equally importantly, a frank appraisal of the limitations of the evidence upon which this therapy recommendation is based. In preparing this review, we queried several large transplantation centers across the country regarding their practices in this disease subset. Responses varied, with highly experienced transplant physicians unified in their uncertainty regarding the best approach. Patients should be made aware of this uncertainty, and historical precedent should not be presented as proven benefit. This is the basis of informed consent.

Contribution: M.T.N. performed literature review; and both authors wrote and edited the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interest.

Correspondence: Mariam T. Nawas, Department of Medicine, Comprehensive Cancer Center of the University of Chicago, 5841 S Maryland Ave, Chicago, IL, 60637; email: nawasm@uchicagomedicine.org.