Sensitivity to ALK-Directed Therapy in Osteosarcoma With an Acquired ALK Rearrangement

Abstract

Background

Sarcomas are clinically, histologically, and molecularly heterogeneous, with at least 70 distinct subtypes occurring in bone and soft tissues.¹ Molecular alterations most commonly involve tumor suppressors (eg, cell cycle control genes, TP53), DNA damage response genes, epigenetic regulators, and, rarely, receptor tyrosine kinases (RTK).^2,3 The prevalence of a given class of alterations varies widely among histologic types. High-grade osteosarcoma is a malignant tumor of bone with diverse genomics, including mutations in TP53, RB1, FGFR1, IGF1R, and DNA structural alterations.^4-7 In most sarcomas, including osteosarcoma, chemotherapy has traditionally been the backbone of therapy as there is an absence of classic kinase alterations as the main drivers (with exceptions such as GI stromal tumor [GIST], dermatofibrosarcoma protuberans, and inflammatory myofibroblastic tumor [IMT]).

Anaplastic lymphoma kinase (ALK) is an RTK and an oncogenic driver in non–small-cell lung cancer (NSCLC), anaplastic large-cell lymphoma, and IMT, among other tumors. Constitutive expression, most commonly through gene fusions, leads to ligand-independent activation and oncogene addiction. This finding has led to the development of multiple tyrosine kinase inhibitors now approved for ALK-rearranged NSCLC, with increasing evidence supporting the use of ALK inhibitors for other ALK-dependent tumor subtypes.^8-11

Herein, we report a 73-year-old man with an acquired ALK-rearranged osteosarcoma. With targeted therapy, the patient showed prolonged disease stabilization of his previously progressing tumor. Phylogenetic analyses revealed a complex clonal relationship among the primary tumor, recurrences, and metastases.

Case

Clinical History

The patient first presented at age 63 years with a 6.6-cm right medial condyle lytic lesion and an associated pathologic fracture (Fig 1). Biopsy revealed high-grade osteosarcoma, and staging studies were free of metastases. He initiated chemotherapy with methotrexate, doxorubicin, and cisplatin. After two cycles, the lesion was resected and found to be 20% necrotic. After surgery, he received adjuvant ifosfamide and etoposide. Six years later, biopsy confirmed recurrence in the soft tissue of the thigh. He underwent resection followed by adjuvant radiation. Unfortunately, within 6 months, he developed lung metastases. Attempts were made to control metastases with ablations, stereotactic body radiation (SBRT), chemotherapy, and resection. At year 7 after first presentation, there was further progression. He was a life-long nonsmoker with no other significant medical history. The longitudinal timeline, including systemic therapy, tumor morphology, and molecular genotype, is shown in Figure 2. Detailed histopathologic images are shown Figure 3.

FIG 1.

Coronal computed tomography image of a lytic lesion centered in the medial femoral condyle corresponding to the primary osteosarcoma at presentation. Note the presence of a pathologic fracture.

FIG 2.

Clinical timeline, tumor morphology, and molecular genotype. Note the lack of overt osteoid morphology coincides with emergence of the ALK rearrangement seen in the recent recurrences. The designations in parenthesis are for comparison to Figure 5. Chemo, chemotherapy; CTMA, CT microwave ablation; Di, diaphragm; empty box, not tested; Kn, knee; neg, negative; pos, positive; RadioTx, radiotherapy; RLL, right lower lobe; RML, right middle lobe; RUL, right upper lobe; Th, thigh.

FIG 3.

Representative histopathologic images are as follows: (A-D) thigh recurrence at 6 years (ALK rearrangement negative) showing (A) chondroblastic features with a chondroid matrix and variable cellularity, (B) telangiectatic features with lakes of blood admixed with malignant cells, (C) prototypical high-grade osteosarcoma with tumor cells associated with osteoid and woven bone, and (D) giant cell–rich morphologies. (E) Right upper lobe metastasis at 7 years (ALK rearrangement–negative) with prototypical high-grade osteosarcomatous morphology. (F) Right middle lobe metastasis at 8 years (ALK rearrangement positive) showing high-grade pleomorphic histology, with pleomorphic tumor cells, and scattered multinucleated giant cells without definitive osteoid or woven bone deposition. Although the thigh recurrence showed a variety of morphologies, the lymphovascular invasion showed the prototypical osteosarcoma morphology as depicted by (C).

As discussed below (molecular profiling), the ALK fusion was first detected in a right middle lobe (RML) lung metastasis, which received microwave ablation and SBRT before resection, unlike previous sites. After further progression, the patient was initiated on off-label alectinib. There was initial tumor growth (+40.5% by Response Evaluation Criteria in Solid Tumors [RECIST] 1.1; Fig 4), followed by disease stabilization on subsequent scans (+26.6% from baseline). After 5 months of alectinib, there was progression and he was transitioned to lorlatinib. Serial imaging confirmed stable disease for over 400 days (best response, stable disease [–11% by RECIST 1.1]; Fig 4). The patient continued lorlatinib with clinical benefit for 14 months, at which point he developed a wound infection in the right knee. This was managed with antibiotics and surgical debridement. The patient was admitted to his local hospital with sepsis. Despite maximal medical intervention, his health condition continued to decline. Ultimately, the decision was made to transition to comfort measures, and he passed away peacefully with his family at the bedside.

FIG 4.

RECIST 1.1 measurements of the lung metastases during ALK-targeted therapy. (A) After initial growth, the target tumors stabilized on alectinib for approximately 10 weeks, before progression. (B) Subsequently, the patient received lorlatinib and had stable disease for nearly 60 weeks. Axial CT image of a lung metastases (circle) (C) before and (D) after 2 months on lorlatinib. (E) cfDNA alteration % change while on lorlatinib. cfDNA, cell-free DNA.

Histology

The primary right femur osteosarcoma was characterized by clusters and sheets of tumor cells associated with osteoid and woven bone, prototypical of a high-grade osteosarcoma. A thigh recurrence at 6 years displayed intermingling of conventional (osteoblastic), chondroblastic, giant cell–rich, and telangiectatic (cystic) areas, representing a recurrent high-grade osteosarcoma. Resection of a right upper lobe (RUL) metastasis at year 7 showed high-grade osteosarcoma with osteoid and woven bone. The resected metastases in the RML, right lower lobe (RLL), and diaphragm at year 8 displayed high-grade pleomorphic histology, with sheets of hyperchromatic tumor cells, extensive necrosis, and scattered multinucleated giant cells. Although there were no overt osteoid or woven bone deposition, that tumor was considered compatible with metastatic/recurrent osteosarcoma because of its histologic similarities to those in the previous specimens, and the subsequent molecular workup. Moreover, there were no distinctive histologic features in these specimens that would suggest to the pathologist to evaluate for an acquired ALK fusion event (Fig 3).

Molecular Genetic Profiling

The initial clinical molecular profiling was performed on the RML metastasis at year 8 after the primary diagnosis (Fig 2). Comprehensive genotyping using RNAseq showed an EML4::ALK fusion (confirmed by immunohistochemistry [IHC] and florescence in situ hybridization [FISH]). In addition, there were TP53^G199V and FGFR1^M276I mutations, as well as CDKN2A loss by DNAseq, the latter also confirmed by FISH. The presence of ALK fusion was confirmed in the concurrent RLL and diaphragm metastases by IHC. The follow-up cytology specimen from RLL a year later had overlapping alterations of ALK, TP53, and CDKN2A, and lacked the FGFR1 mutation.

The primary tumor and thigh recurrence at 6 years were negative for ALK rearrangement by FISH or IHC (next-generation sequencing [NGS] failed because of decalcification). The RUL tumor with classic osteosarcoma morphology at year 7 had overlapping alterations with the previously described RML tumor, including TP53 and FGFR1 mutations, as well as CDKN2A loss, and it also did not harbor the ALK fusion (confirmed by RNAseq, FISH, and IHC).

Clinical circulating cell-free DNA (cfDNA) testing^12,13 was performed at months 1, 6, and 13 after starting lorlatinib. First cfDNA showed the EML4::ALK fusion, TP53^G199V, and FGFR2^S354C. At 6 months, cfDNA showed the EML4:ALK fusion, an ALK^L1196M mutation (known to confer resistance to crizotinib/alectinib), and the TP53^G199V. The FGFR2^S354C was not detectable. At 13 months, cfDNA showed the ALK fusion, the ALK^L1196M resistance mutation, TP53^G199V, FGFR2 ^S354C, and loss of heterozygosity in BRCA2/ATM/PALB2. From 6 through 13 months of lorlatinib, there was stabilization of the ALK/EML4::ALK allele fractions and decrease in the dominant TP53 allele fraction (Fig 4E).

Phylogenetic Analysis

Phylogenetic reconstruction of the tumor evolutionary history showed shared variants in multiple polyguanine repeats in independent samples from the primary knee tumor, thigh recurrence, and metastases to the lung and diaphragm, confirming that all tumors were clonally related. On the phylogenetic tree (Fig 5), samples segregated along two main branches: one branch comprised all the primary tumor samples (Kn1-9) and one region from the thigh recurrence (Th1); the other branch contained the second thigh recurrence region (Th2) and all thoracic metastases. These results are compatible with a scenario in which a localized subclone in the year 6 thigh recurrence seeded the lung metastases. Although the putative metastasis-seeding region of the thigh recurrence (Th2) and RUL metastasis (RUL1-3) at 7 years were negative for the ALK fusion, all other lung metastases (RML1-3, RLL1-2), as well as the diaphragm metastases at 8 years (Di1-3), were positive. This indicates a descendant of the putative metastatic subclone resident in Th2 acquired the ALK fusion, and that some metastases (RLL, RML, and Di) were seeded after this acquisition, while other metastases (RUL) were seeded independently (either earlier in time or by a parallel descendant of the Th2 subclone).

FIG 5.

Polyguanine-based phylogenetic tree (right) showing the clonal relationship between the primary osteosarcoma (Kn1-Kn9), local recurrence (Th1-2), and metastatic RUL tumor (RUL1-3) without the ALK rearrangement, as well as metastatic concurrent RML (RML1-3), RLL (RLL1-2), and Di (Di1-3) tumors (left) with the ALK rearrangement. Di, diaphragm; RML, right middle lobe; RLL, right lower lobe; RUL, right upper lobe.

Autopsy Findings

Autopsy was performed after consent. There was a large RLL metastasis with extensive necrosis (>90%), along with inferior vena cava tumor thrombus. Microscopically, tumor cells appeared spindled-pleomorphic and showed diffuse strong ALK immunoreactivity, consistent with ALK-rearranged high-grade osteosarcoma. Bilateral patchy acute pneumonia and infection of right leg soft tissue were identified. The cause of death was metastatic high-grade osteosarcoma complicated by infection.

Methods are discussed in the Data Supplement (Appendix 1).

Discussion

Here, to our knowledge, we document the first known acquired ALK rearrangement in osteosarcoma. Reconstructing the timeline and integration of morphologic, molecular, and phylogenetic findings confirmed shared clonal origin between the primary and the recurrent/metastatic tumors despite the morphologic and molecular variety. The sensitivity to ALK-targeted therapy was demonstrated with prolonged disease stabilization of the previously rapidly progressing disease and extensive tumor necrosis in the autopsy. Finally, the changes in cfDNA allele fraction and emergence of a resistance mutation confirm antitumor pressure. Although this is a rare occurrence, our report highlights the potential for acquired kinase mutations in heavily treated malignancies and suggest ALK-directed therapy may be an option for acquired ALK rearrangements.

In the present case, the ALK rearrangement first emerged in lung metastases 8 years after resection of the primary. Phylogenetic analysis revealed that a subclone in the patient's thigh recurrence was the likely origin of the lung metastases. Although the patient received multiple cycles of chemotherapy and radiotherapy, interestingly, the first ALK-rearranged tumors developed in the setting of direct tumor exposure to microwave ablation and radiation, which may be suggestive of a thermal and/or radiation-related event or resistance mechanism. Emergence of ALK rearrangements in response to therapy has been reported in lung carcinomas^14,15; however, the chemotherapy/radiotherapy/ablation status of these tumors was not clearly documented. To our knowledge, this phenomenon of an acquired ALK rearrangement in the setting of extensive therapies has not been documented in sarcomas. In addition, there is only a single patient with osteosarcoma in a previous case series of ALK-rearranged nonlung solid tumors.¹⁶ However, the context of primary versus acquired status and the oncogenic role of the ALK rearrangement are unclear on the basis of the data provided. Overall, to our knowledge, the findings of our case represent the first actionable ALK rearrangement in osteosarcomas, as well as its acquired status, which previously has not been established in sarcomas.

In addition to the molecular evolution of this osteosarcoma, there were morphologic changes over time. The resected RML metastasis, the first tumor where we detected the ALK rearrangement, was morphologically distinct (ie, an absence of osteoid formation) from the previous tumors, including an earlier lung metastasis with typical osteosarcoma morphology. This observation raised the possibility of a new primary tumor as a clinical consideration during the patient's care. This was ruled out by integrating phylogenetic assessment into the NGS findings (Figs 2 and 5). Although the phylogenetic evaluation is not a clinical assay, in the future, it may assist in understanding the temporal evaluation of genetic lesions, both for precursor lesions¹⁷ and subsequent metastases.¹⁸ These findings highlight that relying on morphologic features alone would not be adequate to characterize tumors in scenarios such as this case. Overall, our data emphasize the importance of multisite sampling at different stages of disease progression while integrating molecular data to guide therapeutic considerations. This integrative approach will aid the pathology and oncology teams in determining if there is a second independent primary tumor and identify actionable alterations.

In conclusion, this ALK rearrangement in an osteosarcoma is a rare finding that demonstrates acquired RTK alterations are possible in sarcoma. This supports the concept of serial molecular testing through the course of treatment where targeted therapy may be a therapeutic option. Finally, phylogenetic analysis can provide valuable information on the clonal relationship between primary tumors and metastases for morphologically and molecularly challenging cases, which may have implications for treatment decisions.

AUTHOR CONTRIBUTIONS

Conception and design: Zehra Ordulu, Jochen K. Lennerz, Gregory M. Cote

Financial support: Jochen K. Lennerz, Gregory M. Cote

Administrative support: Jochen K. Lennerz

Provision of study materials or patients: Zehra Ordulu, Yin P. Hung, Jochen K. Lennerz

Collection and assembly of data: Zehra Ordulu, Peter Giunta, Wei-Ting Hung, Yin P. Hung, Judit Simon, Florian J. Fintelmann, Jochen K. Lennerz, Gregory M. Cote

Data analysis and interpretation: Zehra Ordulu, Yin P. Hung, Florian J. Fintelmann, Jochen K. Lennerz, Gregory M. Cote

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).