Discussion
Extramedullary AML, also known as myeloid sarcoma, is characterized by the infiltration of blast cells into normal tissues, as confirmed by histological examination.1 Although it can occur at any age, myeloid sarcoma predominantly presents in the pediatric population.8 Myeloid sarcoma may develop either de novo or concomitantly with systemic AML. In some cases, it may precede the development of systemic disease by several months or serve as the initial manifestation of relapsed AML.5 The most commonly affected sites are other hematopoietic organs (such as the liver, spleen, and lymph nodes), followed by skin and gingivae.6 Localized bone involvement, particularly in the temporal bone, is rare and has only been reported in a few case studies. The underreporting of such cases could be a common issue that may be attributed to the diagnostic challenges associated with distinguishing myeloid sarcoma from other conditions, with misdiagnosis rates reported as high as 47% in recent studies.9
In cases similar to the one presented, several factors can contribute to an erroneous diagnosis. Patients with leukemia are prone to developing otologic symptoms, and acute leukemia patients are particularly susceptible to bacterial, fungal, or viral infections affecting the external, middle, or inner ear.10 Less commonly, thrombocytopenic hemorrhage or direct leukemic infiltration can lead to thickening or bleeding in the external auditory canal and tympanic membrane, skin lesions, acute otitis media or externa with effusions, and various types of hearing loss, including sensorineural, conductive, or mixed.11–14 Temporal bone involvement, specifically, may often manifest as a triad of otalgia, hearing loss, and facial paralysis due to facial nerve involvement; additional symptoms may include retroauricular swelling and pain, ear fullness, and vertigo.15–17 Finally, progressive disease can result in extensive bone erosions, permanent damage, and bacterial superinfections, contributing to the development of acute otomastoiditis.11
In the presented case, the patient initially exhibited two of the three symptoms of the aforementioned triad (otalgia and hearing loss), along with fever, ear effusions, and narrowing of the left external auditory canal. Still, these symptoms are non-specific. Considering the prolonged hematologic remission with no signs of systemic disease or central nervous system involvement, a provisional diagnosis of acute otitis media and externa was made.
Unsurprisingly, the subsequent conservative treatment proved ineffective. In some cases, a CT or magnetic resonance imaging (MRI) scan can aid in establishing the diagnosis. Specifically, extramedullary involvement could present as a well-defined, solid mass.8 Imaging features are variable and site-dependent: craniospinal lesions are mostly homogenous and hyperdense on CT, iso- or hypointense on T1 and hyperintense on T2 MRI, with frequent invasion of adjacent bony structures.7,8Nevertheless, these features may be indistinguishable from other, more common malignancies, such as lymphoma.8 Imaging can also demonstrate less specific findings, as was the case in our patient, posing additional diagnostic challenges. Therefore, given a patient’s history, even non-specific opacification should be considered as potential evidence of neoplastic invasion mimicking a benign condition (such as otomastoiditis). Although PET/CT has demonstrated high sensitivity and specificity in identifying myeloid sarcoma,18 it is costly and not widely available. Most importantly, its results are heterogenous when it comes to concurrent bone marrow involvement,18 and interpretation can be even more challenging in cases of extramedullary bone disease and inflammatory complications. As such, imaging features can be insufficient for distinguishing not only between various neoplasms, but also non-malignant disorders.
Hence, a biopsy of the lesion appears to be the only definitive method for confirming a case of isolated extramedullary relapse or de novo myeloid sarcoma. Despite this, invasive procedures are frequently avoided or not feasible owing to certain frequent comorbidities, such as thrombocytopenia. As mentioned, the reported frequency of myeloid sarcoma has varied from 2.5% to 9.1%.2 In contrast to this, a recent PET/CT-based study revealed a prevalence of 22%.18 Likewise, a large-scale study of newly diagnosed AML patients showed an overall incidence of extramedullary disease of 23.7% based on physical examination and imaging.6 Access to potentially affected sites (including the most typical ones, such as liver and spleen) may be challenging or life-threatening, and these are commonly not subjected to biopsy, which likely contributes to the underestimation of prevalence in previous reports, including the cases of temporal bone involvement. Finally, the morphologic appearance of blasts in myeloid sarcoma can vary and depends on the type of AML and characteristics of differentiation, presenting additional challenges. Nevertheless, immunohistochemistry, flow cytometry, and genetic analysis can usually resolve most issues when a tissue biopsy specimen is available.
When present, extramedullary involvement, whether de novo or in the context of relapse, is generally considered a negative prognostic factor.2,19,20 In a study of newly diagnosed AML, CR rate and OS were significantly reduced for patients with EM AML (median OS of 14 months vs. 26.2 months), but event- and relapse-free survival did not differ in patients with and without extramedullary involvement.20 Still, up to 60% of adult patients first present with EM AML during relapse, as was the case in the current report.21 Expectedly, these patients show inferior outcomes when compared to newly-diagnosed cases, with a median overall survival (OS) of 11.6 months and 19.1 months, respectively;21 however, the prognostic significance of EM AML in the context of relapse is still controversial. In a different study investigating extramedullary relapses among AML patients, there was no difference in survival between extramedullary and bone marrow relapse.22 In contrast, another study showed significantly better survival rates in patients with extramedullary relapse (69% at 6 months) when compared to systemic or combined relapses (27% and 8%, respectively).23 Overall, data are limted and show significant heterogeneity and the independent prognostic value of extramedullary disease is still debated. Despite this, the site and pattern of leukemic involvement seem to be important prognostic factors.6,21 For instance, patients with leukemia cutis showed a significantly inferior OS when compared to other sites of disease (median of 5.7 months vs. 21.9 months, respectively).21,24 Interestingly, rare areas of involvement, such as the bone, may be associated with better survival rates.6 Furthermore, a recent review reported that around 22% of temporal bone MS cases may harbor the prognostically favorable t(8;21) translocation.17 Although genetic profiling was unavailable at our center at the time of this case, earlier data had also indicated at existing associations between myeloid sarcoma and the favorable cytogenetic abnormalities t(8;21) and inv(16); still, results are inconsistent, and several studies showed no significant associations with these core-binding factor mutations, and report a varying incidence of other molecular and genetic abnormalities.5,20,21,25 In summary, both the prognostic significance and the genetic profile of MS may be dependent on disease site and other unexplored risk factors, and further studies are needed to investigate this possibility.
Current management of AML in eligible patients involves standard anthracycline-based multi-agent chemotherapy induction, followed by consolidation, with the possible addition of targeted therapy based on risk stratification and mutational profiling.3 There is a lack of prospective clinical trials, and treatment of extramedullary disease, specifically, remains controversial. However, the onset of isolated extramedullary relapse frequently heralds a bone marrow relapse with a mean interval of around 7 months.26 Accordingly, current recommendations for MS are based mainly on existing AML protocols, with or without the addition of radiation therapy and surgery.2,3 In the present case, local treatment was attempted initially. Radiation therapy has shown excellent response rates (91-97%) and local disease control, with a median progression-free survival of 11 months, in one study.27–29 Unsurprisingly, cases of isolated temporal bone AML relapse with prolonged responses after radiation therapy have previously been reported.15 Regardless, survival is poor even among those achieving clinical remission with local therapy, and the majority are likely to relapse.30 In the current case, systemic relapse occurred soon after the end of radiation therapy, possibly due to delayed recognition of the local relapse. Regrettably, management options for relapsed AML and targeted treatment are limited at our center. Despite the late relapse increasing the odds of reinduction with the previously successful regimen, subsequent chemotherapy failed to achieve CR.
Improving management strategies and outcomes in myeloid sarcoma may involve novel therapy and targeting specific molecular alterations, similar to systemic acute myeloid leukemia. For instance, a recent analysis reported a 45% CR rate among EM AML patients treated with venetoclax and hypomethylating agents, and 38% for patients in a relapsed setting.31 However, genetic events underlying extramedullary disease are still not well understood and remain controversial.20 Potentially targetable mutations commonly encountered in AML, such as those of NPM1, FLT3-ITD, IDH, KMT2A, as well as the presence of KIT, TET1, ASXL1, EZH3, SF3B1, NRAS alterations have been associated with extramedullary disease in several recent next-generation sequencing (NGS) studies.20,21In line with this, a study by Ball et al. demonstrated a complete response in three out of four patients treated with IDH inhibitors based on on-site next-generation sequencing of the MS tumor.21 Therefore, biopsy followed by mutational analysis could be crucial for optimizing future therapy of EM AML. Furthermore, significant discordance has been observed between the molecular profiles of the myeloid sarcoma tumors and concurrent bone marrow disease, suggesting the importance of molecular alterations in the pathogenesis of extramedullary disease.21,32 In one study, up to one third of cases showed molecular discordance, with the majority revealing discordance in prognostically important or potentially targetable alterations.32 Importantly, patients presenting with mutational discordance were shown to have an inferior overall survival, underlining the significance of this clonal heterogeneity not only for the pathogenesis, but also prognostication and targetability of extramedullary AML.32 In practice, obtaining sufficient material for cytogenetic and molecular analyses from MS biopsies, often available only in the form of formalin-fixed-paraffin-embedded tissues, can be challenging.33 Utilization of novel NGS-based analyses in clinical practice may overcome this limitation, enabling proper genetic risk assessment and precision-based treatment.33 Nonetheless, due to the low prevalence and varying presentation of myeloid sarcoma, our understanding of optimal management strategies is limited, and more prospective trials are necessary to evaluate the novel treatment approaches for extramedullary AML.