Diagnostic Challenges in Identifying Concurrent Myeloid Sarcoma and High-Grade Prostatic Adenocarcinoma Within the Prostate

Introduction

Myeloid sarcoma is a mass-forming tumor composed of myeloid blasts that occurs outside the bone marrow (ie, extramedullary) and disrupts the tissue architecture. ¹ In patients with acute myeloid leukemia (AML), the incidence of myeloid sarcoma ranges from 2% to 9%, occurring either concurrently with AML or as a relapse. Myeloid sarcoma may rarely precede the onset of AML and can also manifest as a transformation of other myeloid neoplasms, such as myelodysplastic neoplasms and myeloproliferative neoplasms. ² The most commonly affected sites include the skin, soft tissue, lymph nodes, gastrointestinal tract, and bones. ³ Diagnosing myeloid sarcoma can be challenging as approximately 75% of patients with myeloid sarcoma, in the absence of concurrent manifestations of AML, were initially misdiagnosed. ⁴

Myeloid sarcoma of the prostate is extremely rare, with fewer than 30 patients reported in the English-language literature.^5–7 Eight myeloid sarcomas of the prostate have been reported without a prior diagnosis of a hematolymphoid neoplasm.^4,5,8–13 Given its rarity, prostatic myeloid sarcoma poses significant diagnostic challenges and may be overlooked or misdiagnosed, particularly in the absence of a history of a hematolymphoid neoplasm.

Concurrent prostatic myeloid sarcoma and prostatic adenocarcinoma have not been documented in the English-language literature to date. We present the first 2 examples and discuss the associated diagnostic challenges. These represent the 9th and 10th myeloid sarcomas of the prostate in patients without a history of AML or myeloid sarcoma.

Patient Presentation

Patient 1

A 67-year-old man with a medical history significant for heart failure, coronary artery disease (CAD), severe mitral regurgitation, aortic insufficiency, hypertension, hyperlipidemia, type 2 diabetes mellitus, and chronic kidney disease (CKD), presented with paroxysmal nocturnal dyspnea. He was found to be in acute exacerbation of heart failure with concurrent acute kidney injury. Computed tomography of the abdomen and pelvis revealed mild to moderate hydroureteronephrosis, likely due to bilateral distal ureteral obstruction caused by an infiltrative pelvic mass involving and possibly originating from the bladder, prostate, or seminal vesicles. Prostatomegaly was also noted. Subsequent prostate magnetic resonance imaging (MRI) identified a Prostate Imaging Reporting and Data System (PI-RADS) 5 lesion predominantly affecting the left peripheral zone, with involvement of both the transitional and right peripheral zones. The lesion demonstrated extraprostatic extension into the left posterolateral bladder wall, left mesorectum, and bilateral seminal vesicles, contributing to the observed hydroureteronephrosis. There was also involvement of the prostatic urethra. A transperineal prostate biopsy was performed for further evaluation.

Core needle biopsy of the prostate revealed extensive involvement (>95%) by neoplastic cells, with significant effacement of normal prostate architecture (Figure 1A). There were areas predominantly composed of nests, cords, and individual tumor cells infiltrating the prostatic stroma, exhibiting minimal glandular differentiation or poorly formed glands. Scattered signet ring-like tumor cells with cytoplasmic vacuoles compressing the nuclei were also noted (Figure 2A). The overall morphological features were consistent with high-grade prostatic adenocarcinoma, Gleason score 5 + 4 = 9/10 (Grade Group 5). Additionally, there were areas composed of sheets and cords of neoplastic cells, characterized by fine nuclear chromatin, a high nucleus-to-cytoplasm (N:C) ratio, and occasional nucleoli (Figure 2B to D).

OPEN IN VIEWER

Figure 1.

Patient 1. The 2 tumor components were intermixed, making differentiation challenging on hematoxylin and eosin (H&E) staining alone (A). The high-grade prostatic adenocarcinoma was highlighted by pan-keratin AE1.3/CAM5.2 (B) and NKX3.1 (C). In contrast, the hematologic infiltrate was positive for CD45 (D, 100×).

OPEN IN VIEWER

Figure 2.

Patient 1. High-grade prostatic adenocarcinoma component with scattered signet ring-like tumor cells (A). Additionally, there were regions occupied by myeloid sarcoma largely composed of sheets and cords of neoplastic cells with fine nuclear chromatin, a high nucleus-to-cytoplasm ratio (B, 200×; C, 400×; D, 1000×), and occasional nucleoli (arrow, D). The myeloid blasts demonstrated strong, diffuse lysozyme staining (E, 200×) and cytoplasmic expression of NPM1 (F, 400×).

Notably, these 2 tumor components were intermixed (Figure 1), making it challenging to distinguish them on hematoxylin and eosin (H&E) staining alone. Immunohistochemical (IHC) staining was employed for further characterization. The high-grade prostatic adenocarcinoma component was highlighted by pan-keratin AE1.3/CAM5.2 (Figure 1B), NKX3.1 (Figure 1C), and prostate-specific membrane antigen (PSMA). Additionally, a prominent infiltrate of hematologic cells was identified, which was CD45+ (Figure 1D), lysozyme+ (strong and diffuse staining; Figure 2E), CD68+, CD34−, CD117+/−, CD4+/−, CD7−, and CD56−/+ and negative for CD19, CD20, CD10, BCL6, MUM1, and CD30, with a Ki-67 proliferation index of 20% to 30%. CD21+ follicular dendritic cells were absent, and a few scattered small T cells (CD3+, CD5+) were observed. Chromogenic in situ hybridization for Epstein–Barr virus-encoded RNA was negative. These findings were consistent with a diagnosis of myeloid sarcoma with monocytic differentiation (extramedullary equivalent of AML). Many of the neoplastic cells were also positive for NPM1 using a mutation-specific antibody (Figure 2F), suspicious for an underlying NPM1 mutation.

Overall, the prostate was involved by 2 distinct tumors, with approximately one-third of the neoplastic cells representing high-grade prostatic adenocarcinoma and the remaining two-thirds consisting of myeloid sarcoma. The complete blood count revealed a decrease in hemoglobin at 10.7 g/dL, hematocrit at 34.2%, and mean corpuscular volume at 71.5 fL, with an elevated red cell distribution width at 22.1%. White blood cell count was normal at 8.92 K/µL, as was the platelet count at 273 K/µL. The automated differential showed an increase in neutrophils (74.3%) and immature granulocytes (1.3%), a decrease in lymphocytes (15.6%), and normal levels of monocytes (7.7%), eosinophils (0.3%), and basophils (0.8%).

A subsequent bone marrow biopsy revealed a cellularity of approximately 90%, with a markedly increased myeloid to erythroid ratio. The biopsy showed primitive cells consistent with blasts arranged in sheets, comprising approximately 80% of the total cellularity. IHC staining demonstrated that the blasts were CD34−, CD117+, PU1+, and NPM1+. Flow cytometry of the bone marrow identified 10% myeloid blasts (CD45 dim, CD117+, CD34−/+, CD33+/−, CD13−, HLA-DR−/+, CD64−/+, CD56−/+, CD123+, CD14−, CD19−, CD10−, CD3−). Cytogenetic analysis revealed a normal male karyotype. Rapid heme next-generation sequencing detected several pathogenic mutations commonly associated with AML, including NPM1 (p.W288Cfs*12), ASXL1 (p.P840Lfs*2), SRSF2 I (p.P95H), FLT3-TKD (p.D835Y), and IDH2 (p.R140Q). The final classification was AML with mutated NPM1 (World Health Organization [WHO], revised 4th ed. and International Consensus Classification) and AML with NPM1 mutation (WHO, 5th ed.).

During hospitalization, the patient was evaluated for CAD and heart failure, which revealed no options for catheter-based or surgical repair of his CAD and valvular diseases. Evaluation for CKD showed progressive disease, with the patient deemed not suitable for dialysis. The patient initially started treatment with bicalutamide for prostate cancer but discontinued it a few weeks later due to side effects. Enasidenib, an IDH2 inhibitor, was planned for the treatment of his AML. However, the patient was considered ineligible for aggressive therapies that might prolong life. He passed away 44 days after the initial diagnosis due to complications arising from multiple comorbidities.

Patient 2

A 76-year-old man with a history of renal cell carcinoma, follicular lymphoma (diagnosed in 2017), severe CKD on dialysis, chronic obstructive pulmonary disease (COPD), and type 2 diabetes mellitus presented with an elevated prostatic-specific antigen (PSA) level of 10.48 ng/mL. A prostate MRI revealed a 1.7 cm PI-RADS 5 lesion in the right medial peripheral zone at the base of the prostate, with local extension to the right seminal vesicle and possible extension into the left side of the prostate. A transrectal prostate biopsy performed at an outside institution demonstrated high-grade prostatic adenocarcinoma, Gleason score 4 + 4 = 8 (Grade Group 4), involving the right base and mid-prostate. Subsequent radical prostatectomy at the same institution confirmed prostatic adenocarcinoma with a Gleason score of 4 + 3 and a tertiary pattern 5 (<5%), involving approximately 10% of the prostate, with extraprostatic extension in the right base and invasion of the right seminal vesicle (pT3b N0). In addition, the prostate exhibited a patchy stromal infiltrate of discohesive, medium-sized cells with oval, irregular, or indented nuclei, finely stippled chromatin, and scant cytoplasm. These cells were intermixed with prostatic adenocarcinoma (Figure 3A to C) and surrounded benign prostatic glands (Figure 3D). Similar cellular infiltrates were observed in the majority of the 6 examined pelvic lymph nodes, with atypical cells arranged in sheets and bands (Figure 3E). IHC staining revealed that the atypical cells were CD45 weak+, lysozyme+ (diffuse and strong, Figure 3F), CD68 weak+, myeloperoxidase (MPO)+/− (Figure 3G), CD34−, CD117+/−, CD79−, PAX5−, CD3−, CD4 weak+, CD5−, CD10−, CD15 weak+/−, CD30−, CD43+, CD56+/−, BCL2 weak+, BCL6−, CD21−, cyclin D1−, SOX10−, S100−, CD123+, and synaptophysin- and pan-keratin AE1/AE3−. Ki-67 highlighted approximately 40% of the atypical cells. Neoplastic cells also showed positivity for a mutation-specific antibody for NPM1 (Figure 3H), while p53 displayed a wild-type staining pattern. These findings were consistent with myeloid sarcoma with monocytic differentiation, with immunophenotypic evidence of an NPM1 mutation.

OPEN IN VIEWER

Figure 3.

Patient 2. The radical prostatectomy specimen displayed a stromal infiltrate of discohesive, medium-sized myeloid blasts with oval, irregular, or indented nuclei, finely stippled chromatin, and scant cytoplasm (A-E). Myeloid blasts (A, 200×; arrows, B and C, 400×) were intermixed with prostatic adenocarcinoma (A; arrowheads, B and C) and surrounded benign prostatic glands (D, 400×). Myeloid blasts arranged in sheets also diffusely infiltrated the pelvic lymph nodes (E, 400×) and were lysozyme+ (diffuse and strong; F, 200×) and MPO+/− (G, 200×), with cytoplasmic expression of NPM1 (H, 200×).

The patient opted for comfort care, and no bone marrow biopsy or further treatments were performed. The patient passed away from respiratory failure secondary to exacerbated COPD in the context of multiple comorbidities, 2 months after being diagnosed with myeloid sarcoma.

Discussion

Myeloid sarcoma in the prostate is exceptionally rare, with fewer than 30 patients reported in the literature. Of these, 8 have been documented as the first presentation, without a prior history of hematolymphoid neoplasms. To date, there have been no reports in the English-language literature of concurrent myeloid sarcoma and prostate cancer within the prostate. Herein, we present the first 2 such cases identified in a prostate core needle biopsy and radical prostatectomy, highlighting the associated diagnostic challenges.

Hematolymphoid neoplasms that occur alongside nonhematologic neoplasms pose significant diagnostic challenges. They can resemble chronic inflammation, including immune responses, making them easy to misinterpret as reactive changes to a solid organ neoplasm, particularly when they are low grade. The prostate is the third most common site for collision tumors involving hematolymphoid neoplasms (12%), after colon (17%) and breast (15%). The most frequently identified hematolymphoid neoplasms across all sites include chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL; 18%), diffuse large B-cell lymphoma (17%), follicular lymphoma (14%), marginal zone lymphoma (10%), and myeloid sarcoma/AML (8%). ¹⁴ Among 12 hematolymphoid neoplasms concurrent with prostate cancer, there were 6 CLL/SLL, 2 follicular lymphomas, one peripheral T-cell lymphoma, one non-Hodgkin lymphoma, one multiple myeloma, and one hairy cell leukemia. CLL/SLL was initially missed in 2 specimens, one in a prostate core needle biopsy, and the other in a lymph node dissection. Both diagnoses were amended after peripheral blood flow cytometry was performed. ¹⁴ Therefore, it is crucial to consider the possibility of collision hematolymphoid neoplasms before assuming that lymphoid infiltrates are non-neoplastic.

None of the previously reported 8 myeloid sarcomas in the prostate, in the absence of prior hematolymphoid neoplasms, were concurrent with prostate cancer (Table 1). The median age of patients was 66 years (range: 32-73). Half of the myeloid sarcomas were identified on prostate biopsies and half on transurethral resection of the prostate specimens. AML was confirmed through bone marrow biopsies and ancillary tests in 7 of the 8 patients, with one showing normal bone marrow. Chemotherapy was initiated in all patients, and radiotherapy to the prostate was administered in 2 patients. In a median follow-up of 8.8 months (range: 0.8-24), half of the patients died of disease, with a median survival time of 4.3 months (range: 0.8-24) following the initial diagnosis, while 3 achieved complete remission. Follow-up data were unavailable for one patient. One patient was initially misdiagnosed with lymphoma and treated accordingly. Upon relapse of AML, remission-induction chemotherapy was initiated, leading to complete remission.

OPEN IN VIEWER

Table 1.

Summary of Myeloid Sarcoma in the Prostate as First Presentation, Without a Prior History of AML or Myeloid Sarcoma.

Patient	References	Age (y)	Prostate specimen	Concurrent prostate cancer	Subsequent BM examinations	Treatment	Follow-up (m)	Outcome
1	Frame 8	61	TURP	No	AML	Chemo	5.5	DOD
2	Quien 9	67	TURP	No	AML with maturation	Chemo a , RT to the prostate	19	CR
3	Spethmann 4	32	Biopsy	No	Acute monocytic leukemia	Chemo	3	DOD
4	Tsuboi 10	71	Biopsy	No	AML	Chemo	NA	CR
5	Mneimneh 5	65	TURP	No	AML with inv(16)(p13.1;q22)	Chemo	0.8	DOD
6	Nitta 11	71	Biopsy	No	AML with maturation	Chemo, RT to the prostate	24	DOD
7	Horvat 12	44	Biopsy	No	AML	Chemo, BMT	NA	NA
8	Nguyen 13	73	TURP	No	Normal	Chemo	12	CR
9	Present patient 1	67	Biopsy	Yes (Gleason score 5 + 4)	AML with NPM1 mutation	Bicalutamide for prostate cancer	1.5	DOD
10	Present patient 2	76	RP	Yes (Gleason score 4 + 3, with a tertiary pattern 5)	Not performed	RP alone	2	DOO

Abbreviations: AML, acute myeloid leukemia; BM, bone marrow; BMT, bone marrow transplantation; Chemo, chemotherapy; CR, complete remission; DOD, dead of disease; DOO, dead of other causes; m, months; NA, not available; RT, radiotherapy; RP, radical prostatectomy; TURP, transurethral resection of the prostate; y, years.

^a

The patient was initially misdiagnosed with lymphoma and treated accordingly. Upon relapse of AML, remission-induction chemotherapy was initiated, leading to a complete remission.

In contrast, we report the first 2 myeloid sarcomas in the prostate concurrent with high-grade prostatic adenocarcinoma in the absence of prior AML or myeloid sarcoma, both of which harbored the NPM1 mutation. Neither patient received treatment for myeloid sarcoma/AML due to multiple comorbidities. One patient died of the disease 44 days after diagnosis, while the other died of respiratory failure in the context of multiple comorbidities, 2 months after being diagnosed with myeloid sarcoma. This is consistent with the poor prognosis observed in the 8 previously reported myeloid sarcomas. The prevalence of NPM1 mutation in myeloid sarcoma has been reported to range from 10% to 30%.^15–18 NPM1 frameshift mutations, typically caused by a 4 base pair insertion in exon 12 as observed in patient 1, alters the C-terminal DNA-binding domain of the protein and results in aberrant nuclear export and cytoplasmic localization. This mislocalization is thought to be critical for its role in leukemogenesis. ¹⁹ Although studies showed that NPM1 expression was higher in prostate cancer compared to benign prostatic tissue and promoted prostate cancer progression,^20,21 NPM1 frameshift mutations have not been reported in prostate cancer. Furthermore, cytoplasmic expression of NPM1 was not observed in the prostatic adenocarcinoma component in patient 1 using a mutation-specific antibody. Therefore, it is likely that the concurrent myeloid sarcoma and prostatic adenocarcinoma were not clonally related. The presence of NPM1 mutations in both patients in the present study may be coincidental. No prior examples of myeloid sarcoma with NPM1 mutation have been reported in the prostate, although NPM1 IHC may not have been available in earlier studies. Further research is warranted to elucidate the significance of these findings.

The presence of myeloid sarcoma at the time of a diagnosis of AML does not appear to impact prognosis significantly. ²² However, patients with concurrent myeloid sarcoma and AML that are molecularly discordant tend to have a worse prognosis compared to those with molecularly concordant disease. ²³ Patients presenting with isolated myeloid sarcoma, such as patient 8 in Table 1, generally have a more favorable prognosis. ²⁴

The primary differential diagnoses of prostatic myeloid sarcoma include high-grade prostatic adenocarcinoma, neuroendocrine differentiation of prostatic adenocarcinoma, such as small cell neuroendocrine carcinoma (SCNEC), small round blue cell tumors, and other hematolymphoid neoplasms like lymphomas and blastic plasmacytoid dendritic cell neoplasm. ²⁵

High-grade prostatic adenocarcinoma, particularly Gleason pattern 5 as observed in both patients in the current study, is characterized by minimal glandular differentiation, sheets or cords of tumor cells, and single cells. This can make it challenging to distinguish from myeloid sarcoma on H&E staining. Both may exhibit discohesive, and poorly differentiated morphology, with myeloid sarcoma occasionally displaying nucleoli, which are commonly seen in prostate cancer. IHC using markers for prostatic adenocarcinoma, such as PSA and NKX3.1, alongside markers for myeloid sarcoma, including CD45, CD68, lysozyme, CD117, and CD34, may aid in differentiating these 2 entities. However, when both populations are intermixed as shown in the present lesion, identifying the presence of a second neoplastic population can be challenging, potentially leading to the omission of appropriate IHC stains. Therefore, the threshold for initiating IHC studies should be kept low in such instances to avoid misdiagnosis in these challenging diagnoses.

De novo SCNEC of the prostate is extremely rare, accounting for approximately 0.03% of prostate cancers. ²⁶ However, it can also develop in 15% to 20% of patients during later stages of prostate cancer progression, particularly those treated with hormone therapies. ²⁷ The pathogenesis of prostatic SCNEC is not fully understood, but molecular data suggest that many SCNECs may arise from transdifferentiation of conventional adenocarcinoma, even in treatment-naïve patients. ERG gene rearrangements have been shown to be concordant between the adenocarcinoma and SCNEC components. ²⁸ Loss of the RB1 and TP53 tumor suppressor genes are key genomic drivers in most patients, like SCNEC at other sites. ²⁹ Neuroendocrine markers such as synaptophysin and chromogranin A can be useful for diagnosis, although some SCNECs may completely lack labeling for these markers.

Additionally, small round blue cell tumors, such as poorly differentiated synovial sarcoma, Ewing sarcoma, and desmoplastic small round cell tumor, should be differentiated from myeloid sarcoma in the prostate. Primary synovial sarcoma can rarely present in the prostate. ³⁰ Poorly differentiated synovial sarcoma, characterized by sheets of undifferentiated tumor cells with a high N:C ratio and scant cytoplasm, may resemble myeloid sarcoma. Primary Ewing sarcoma and desmoplastic small round cell tumor of the prostate are exceedingly rare, with only a few reported patients.^31–34 Molecular testing for gene fusions is crucial for establishing an accurate diagnosis in these tumors.

Other hematolymphoid neoplasms, such as lymphomas and blastic plasmacytoid dendritic cell neoplasm, should also be distinguished from myeloid sarcoma in the prostate. IHC panels can aid in differentiating these entities. For instance, blastic plasmacytoid dendritic cell neoplasm typically shows positive staining for CD123, CD4, and CD56, and negative staining for lysozyme, CD117, and MPO. ²⁵ In the present study, the diffuse and strong expression of lysozyme excluded blastic plasmacytoid dendritic cell neoplasm.

In conclusion, we present the first 2 examples of concurrent high-grade prostatic adenocarcinoma and myeloid sarcoma within the prostate, highlighting the importance of maintaining a high index of suspicion, particularly those with high-grade morphology.