Lung Cancer

Durable response to serial tyrosine kinase inhibitors (TKIs) in an adolescent
with metastatic TFG-ROS1 fusion positive Inflammatory Myofibroblastic
Tumor (IMT)
Katrina M. Ingley a
, Debbie Hughes b
, Michael Hubank c
, Daniel Lindsay d
, Andrew Plumb e
,
Rachel Cox f
, Louis Chesler g
, Sandra J. Strauss a,h,
*
a London Sarcoma Service, University College London Hospitals NHS Trust, United Kingdom b Clinical Genomics Translational Research, Centre for Molecular Pathology, The Institute of Cancer Research Sutton, United Kingdom c Molecular Diagnostics Department, The Institute of Cancer Research and Clinical Genomics, The Royal Marsden NHS Foundation, London, United Kingdom d Department of Histopathology, Royal National Orthopaedic Hospital NHS Trust, Stanmore, United Kingdom e Department of Radiology, University College London Hospitals NHS Trust, United Kingdom f Department of Paediatric Oncology, Bristol Royal Hospital for Children, Upper Maudlin Street, Bristol, BS2 8BJ, United Kingdom g Paediatric Tumor Biology, Division of Clinical Studies, The Institute of Cancer Research, London, United Kingdom h UCL Cancer Institute, London, United Kingdom
ARTICLE INFO
Keywords:
Inflammatory myofibroblastic tumor (IMT)
ABSTRACT
Here, we present the case of an adolescent with a rare metastatic Inflammatory myofibroblastic tumor (IMT)
harboring a TFG-ROS1 fusion initially detected on tumor progression and retrospectively identified in the pri￾mary tumor after targeted RNA sequencing. The patient benefitted from sequential TKIs over a 5-year period
with response to the third generation ALK/ROS inhibitor, lorlatinib leading to resection of the primary tumor.
Detailed molecular analysis can identify targetable oncogenic kinase fusions that alters management in patients
with unresectable disease and should be considered in all patients.
1. Introduction
Inflammatory Myofibroblastic Tumors (IMTs) are rare mesenchymal
tumors with a variable natural history that affect patients of all ages
including children and adolescents. IMT is often a diagnostic challenge
due to a wide and varied morphologic spectrum [1]. Targeted Next
Generation Sequencing (NGS) has shown that up to 85 % of IMT are
driven by kinase fusions, the majority involve ALK (~60 %) and other
gene fusions, ROS1 (~10 %) and PDGFRß (~3%) [2]. ROS1 and ALK
receptor tyrosine kinase domains share homology in amino acid
sequencing and structurally related adenosine triphosphate (ATP)
binding sites [3]. Small molecule TKIs with dual inhibitory activity
against ALK and ROS1 have demonstrated efficacy across ALK and
ROS1- rearranged human cancers, most notably in NSCLC [3].
ALK and ROS1-rearranged cancers acquire resistance to TKIs leading
to disease progression. Second and third generation selective ALK/
ROS1 inhibitors are able to overcome this resistance with inhibitors such
as Lorlatinib (Pfizer) demonstrating a 42 % response rate in previously
treated ALK- and ROS1-rearranged NSCLC and are highly active in the
CNS and against TKI-resistant mutants [4]. There is no standard of care
for management of patients with unresectable IMT. CREATE, a
biomarker driven study conducted by the EORTC is the only prospective
phase 2 trial conducted in IMT and demonstrated crizotinib to be highly
active in ALK-positive tumors, as well as achieving disease control for
ALK-negative patients [5]. Recent case reports on patients with
ALK-rearranged IMT have demonstrated partial response to lorlatinib,
leading to complete resection for one patient [6,7]. Here, we describe
the first case of an adolescent with a metastatic ROS1 –rearranged IMT
treated with multiple ALK/ ROS1 inhibitors including lorlatinib
demonstrating a durable response that led to resection of the primary
Abbreviations: IMT, Inflammatory myofibroblastic tumor; TKIs, tyrosine kinase inhibitors; NGS, next generation sequencing; ATP, adenosine triphosphate; NSCLC,
non-small cell lung cancer; FISH, fluorescent in-situ hybridization; IHC, immunohistochemistry; EORTC, European Organisation for Research and Treatment of
Cancer; DNA, deoxyribonucleic acid; RNA, ribonucleic acid; FFPE, formalin-fixed paraffin-embedded; EpSSG, European pediatric Soft Tissue Sarcoma Study Group.
* Corresponding author at: University College London Hospitals NHS Trust, 250 Euston Road, London, NW1 2PG, United Kingdom.
E-mail address: [email protected] (S.J. Strauss).
Contents lists available at ScienceDirect
Lung Cancer
journal homepage: www.elsevier.com/locate/lungcan

https://doi.org/10.1016/j.lungcan.2021.05.024

Received 18 April 2021; Received in revised form 15 May 2021; Accepted 22 May 2021
Lung Cancer 158 (2021) 151–155
152
tumor.
2. Case presentation
A 14-year-old female presented with a 2-month history of neuro￾pathic pain centred on the right scapula, radiating into the axilla and
arm with a profound impact on quality of life. An Xray then magnetic
resonance imaging (MRI) revealed a heterogeneous 9- cm mass located
in the right upper lobe (RUL) of the lung extending to the lung apex,
pleura and mediastinum and staging revealed multiple enhancing
cortical brain lesions (Fig. 1A and B). Open biopsy of the right frontal
brain metastasis, and three computed tomography (CT)- guided biopsies
of the primary thoracic mass with IHC and FISH for ALK led to the
diagnosis of metastatic ALK-negative IMT following several expert
Fig. 1. a–j: Timeline depicting the clinical
course and radiological response to treat￾ment.
Baseline axial T1 weighted MRI of the brain
following intravenous gadolinium (A) and axial
T2 weighted MRI of the thorax show multiple
brain metastases [arrows in (A)] and a bulky
mass in the RUL of the lung [measured in (B)].
Following 8 cycles of crizotinib and excision
biopsy of the right frontal brain metastasis, only
a small rim of dural enhancement remains in
the brain [arrow in (C)], and the RUL mass has
reduced in size [measured in (D)], with re￾aeration of the anterior RUL [arrow in (D)].
Although the brain imaging remained stable
[arrowed in (E)], intrathoracic recurrence
despite treatment with brigatinib was diag￾nosed based on the enlarging RUL mass which
was beginning to invade into the right chest
wall and brachial plexus [arrowed in (F)].
Response is maintained in the brain, shown in
(G), with intrathoracic imaging (H) after 5 cy￾cles of lorlatinib showing good partial response
in the RUL mass and restoration of the extrap￾leural fat planes, with retraction of the mass
away from the right brachial plexus. After a
total of 11 cycles of lorlatinib, intracranial
response was maintained with only a trace of
residual dural enhancement [arrowed in (I)]. Further response in the chest is shown in (J), with a clear rim of fat between the RUL tumor and the subclavian vessels
and brachial plexus, deeming the disease resectable.
Fig. 2. a–f: Pathology and response to
treatment.
2A. VATS biopsy, RUL, February 2016 (H&E
x10). A tumor with a vaguely fasciculated ar￾chitecture composed of elongated spindle cells
with mildly atypical spindle shaped nuclei and
amphophilic cytoplasm imparting a myofibro￾blastic morphology. The tumor cells are
embedded in a collagenous stroma. There is an
associated infiltrate of chronic inflammatory
cells composed predominantly of lymphocytes
with occasional plasma cells. Note how the
tumor cells have a loose fascicular growth with
spindle cells featuring elongated cytoplasmic
processes, a phenotypic feature associated with
ROS1 rearranged IMT [10]. Very occasional
mitotic figures were present (<1/10 high power
fields (HPF)) and there was no necrosis. 2B.
VATS biopsy, RUL, February 2016 (H&E x20).
Higher power image of the lung tumor, further
demonstrating the features described in 2A.
Note there are scattered larger cells with a
rounded morphology and conspicuous eosino￾philic nucleoli (within box). 2C. Primary tumor
from lung x20 (ALK IHC). Immunohistochem￾istry for ALK1 (clone 5A4) is negative in the
tumor cells. The spindle cells demonstrated low proliferative activity. 2D. Needle biopsy from right supraclavicular fossa mass, October 2019, (H&E x20). Pathology
confirms tumor with similar morphological features as described in 2A and B. Note the degree of cytological atypia- greater variation in nuclear size, hyperchromasia
and mitotic activity compared to February 2016. This tumor also showed areas of coagulative necrosis (not pictured). 2E. (H&E x4) and 2 F. (H&E x10): Resection
from RUL/ chest wall December 2020, post lorlatinib treatment. The tumor has shown an excellent response to therapy, with diffuse areas of metaplastic ossification,
calcification and fibrosis. Only approximately 5% residual viable tumor was identified on the resection specimen.
K.M. Ingley et al.
Lung Cancer 158 (2021) 151–155
153
pathology opinions (Fig. 2A–C).
The patient received minimal benefit from cytotoxic chemotherapy
including high-dose methylprednisolone and mycophenolate mofetil
(MMF), methotrexate and vinblastine and ifosfamide. Neuropathic pain
and fatigue escalated rendering her housebound. Eighteen months post
diagnosis the patient commenced crizotinib (250 mg twice daily)
through entry to the CREATE study [5]. Immunohistochemistry, FISH
and RNA analysis using the Archer CTL fusion panel performed on study
was negative for ALK and ROS1 [5]. The patient achieved a rapid and
excellent clinical response to crizotinib with resolution of all symptoms
and near complete resolution of brain metastases over 18 months
(Fig. 1C and D). The primary tumor however remained stable by RECIST
1.1 criteria and despite ongoing clinical benefit, CT and MRI imaging
following 22 cycles demonstrated small volume new pleural disease,
consistent with RECIST progressive disease and the patient was with￾drawn from the study, quickly becoming symptomatic once again.
Despite multiple biopsies it was not possible to obtain sufficient tumor
cells for further genetic analysis.
The patient had a rapid clinical response to the second generation
ALK/ ROS1 inhibitor brigatinib (180 mg daily) obtained on compas￾sionate access for 9 months before further disease progression (Figs. 1E
and F, 2 D). Biopsy of the primary thoracic tumor at this time detected a
TFG-ROS1 fusion through RNA sequencing performed via entry to an
ongoing paediatric sequencing study, CRUK-Stratified Medicine Paedi￾atrics study, SMPaeds. Third generation lorlatinib (100 mg daily) was
initiated through a compassionate access scheme with a good partial
response (Fig. 1G and H). The patient completed 1-year of lorlatinib and
was able to undergo complete surgical resection of residual disease 5-
years from diagnosis Fig. 2E and F).
3. Methods
3.1. Molecular analysis
Patient informed consent for molecular analysis and related research
was obtained through the UCL/UCLH biobank for studying health and
disease (National Research Ethics Committee reference 15/YH/0311)
and CRUK Stratified Medicine Paediatrics (ISRCTN21731605). A biopsy
from the primary thoracic tumor at diagnosis and relapse were evalu￾ated by NGS (see Molecular methods, Appendix 1). Tumor content was
30 % and 90 %, respectively. Of significance, the primary tumor sample
had appreciable reactive changes and crush artefact hindering accurate
estimation of tumor content.
Multiple DNA sequencing approaches in the DNA derived from FFPE
from the primary and relapse biopsies and blood for germline analysis
failed to identify any clinically significant variants associated with the
IMT phenotype (Fig. 3). RNA analysis, using the TruSight RNA Pan￾Cancer Panel and RNA-Sequencing Alignment App v2.0.1. (Illumina
Inc., San Diego, CA, USA) identified a TFG-ROS1 fusion in the relapse
biopsy that was retrospectively confirmed in the primary biopsy (Fig. 4).
4. Discussion
This patient required repeat biopsies and expert pathology review to
Fig. 3. Clinical course and genomic profiling.
Summary of all analyses performed on the primary, relapsed tissue samples and germline DNA to find the causative variant. DNA sequencing employed a clinically
validated custom panel targeting 92 genes associated with paediatric cancers (473 kb) and in the relapse sample, an exome panel targeting 19,396 genes (39Mb).
DNA analysis was designed to detect single nucleotide, copy number and structural variants. Panel sequencing of germline DNA identified a SNV, BCOR c.-34T > C, at
a variant allele frequency (VAF) of 56 %, confirmed in both the primary (VAF 35 %) and the relapse (VAF 21 %) tumor but deemed to be of no clinical significance as
it is unlikely to affect splicing, has not been reported previously and is not associated with the IMT phenotype. No clinically actionable, somatic variants, defined as
those that alter gene function, confer drug resistance or influence disease prognosis or diagnosis, were identified in the panel data or genes commonly associated with
resistance to TKI inhibition including PDGFRA, RET, ROS1 in the WES data.
K.M. Ingley et al.
Lung Cancer 158 (2021) 151–155
154
confirm a morphological diagnosis of IMT, with IHC and FISH negative
for ALK. Targeted RNA sequencing at disease progression identified a
TFG-ROS1 fusion that was discordant with the previous negatively re￾ported FISH for ROS1. Challenges in accurate diagnosis of IMT are well
described with a European pediatric Soft Tissue Sarcoma Study Group
(EpSSG) prospective trial demonstrating 20 patients (25 %) with an
initial diagnosis of IMT having their diagnosis amended after central
review. [8] In addition, molecular diagnostics are not routinely avail￾able or standardised for this rare malignancy. Chang et al. demonstrated
in a cohort of 33 patients with thoracic IMT that by using an array of
molecular techniques an oncogenic tyrosine kinase fusion protein could
be identified in every patient, including 6 ROS1-rearranged IMTs, 1 of
which did not stain for ROS1 by IHC [9]. Similarly, they report an
overall sensitivity for FISH at 86 %, with 4/30 thoracic IMT cases (13 %)
negative by FISH that had a fusion confirmed with targeted RNA
sequencing, including one case of TFG-ROS1 [9].
Here, we demonstrated the utility of targeted RNA sequencing to
detect therapeutically actionable kinase fusions in a patient with unre￾sectable IMT. It was challenging to identify the kinase fusion in the
primary sample, that was initially detected in the relapse sample. This is
most likely due, at least in part to lower tumor content in the primary
sample; tumor heterogeneity, the lability of RNA in archival FFPE, low
expression levels of fusion transcripts and limitations in testing sensi￾tivity and PCR amplification bias may also be contributing factors. These
potential limitations in diagnostic testing lend support to repeat mo￾lecular re-evaluation at progression for any fusion negative IMT’s.
Disease in this patient was controlled using multigenerational TKIs,
but with each TKI, resistance eventually developed. ALK-negative IMT
may be less responsive to crizotinib and acquire mutations within the
ROS1 kinase domain more frequently than ALK-positive IMT [4,5].
Despite candidate gene analysis, the resistance mechanism was not
identified in this study. Escalation in potency of sequential generation
TKI’s to the highly effective and CNS penetrant lorlatinib established an
objective response that deemed the tumor resectable.
5. Conclusion
This case study demonstrates the efficacy of serial TKIs in a patient
with a very rare tumor, ROS1-rearranged IMT. In patients with ALK￾negative IMTs, detailed molecular analysis at diagnosis and on a
repeat biopsy at progression should be considered to detect rare gene
rearrangements and optimise ongoing therapy.
Funding disclosure
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
Author statement file
Katrina Ingley (KMI): Visualisation; Data curation; Project adminis￾tration; Original draft. Debbie Hughes (DH): Methodology; Formal
analysis; Validation; Writing- review & editing. Michael Hubank (MH):
Methodology, Formal analysis; Validation. Daniel Lindsay (DL): Inves￾tigation, Writing- review and editing. Andrew Plumb (AP): Investiga￾tion, Writing- review and editing. Rachel Cox (RC): Investigation. Louis
Chesler (LC): Resources, ALK inhibitor Writing- review and editing. Sandra Strauss
(SJS): Conceptualisation; Visualisation; Supervision; Writing- review &
editing.
Declaration of Competing Interest
The authors report no declarations of interest.
Fig. 4. TFG Exon 4 is fused to ROS1 Exon 35 in the primary and relapse biopsies.
Primary and relapsed RNA samples showed a TFG:ROS1 fusion (Breakpoint chr3:100,447,701:chr6:117,642,557 as identified by arrows. In all instances paired end
sequencing (2 × 75bp) generated a minimum of 3 million unique aligned reads per sample, in line with supplier recommendations. Analysis software failed to
identify the fusion in the primary biopsy for both TruSight RNA Pan-Cancer and Tru-Sight Tumor 170 panel data; fusion calls are calculated as a weighted average of
individual features including percentage of fusion supporting reads, read counts across fusion breakpoint, alignment qualities and additional quality metrics fusion.
The fusion did not meet the confidence score criteria but is visible when inspecting the data in the genome browser Integrative Genome Viewer at the breakpoint. All
RNA sequencing was performed to diagnostic/ISO standards.
K.M. Ingley et al.
Lung Cancer 158 (2021) 151–155
155
Acknowledgements
We would like to extend our sincere thanks and gratitude to the
patient and her family. Other contributions: The Institute of Cancer
Research and Cancer Research UK (CRUK)- Stratified Medicine Paedi￾atrics Team (SMPaeds); Alison Headford and Dr Elise Gradhand from the
Department of Pathology, Bristol Royal Hospital for Children, Bristol,
UK and Pfizer Oncology UK. MH and the NIHR Centre for Molecular
Pathology are supported by the Royal Marsden Biomedical Research
Centre (BRC). SJS is supported in part by the NIHR UCLH Biomedical
Research Centre. We thank Prof Schoffski, ¨ Dr Wozniak and the EORTC
for confirmation of the molecular analysis performed as part of the
CREATE study.
Appendix A. Supplementary data
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.lungcan.2021.05.024.
References
[1] C.M. Coffin, J.L. Hornick, C.D. Fletcher, Inflammatory myofibroblastic tumor:
comparison of clinicopathologic, histologic, and immunohistochemical features
including ALK expression in atypical and aggressive cases, Am. J. Surg. Pathol. 31
(2007) 509–520.
[2] C.M. Lovly, A. Gupta, D. Lipson, et al., Inflammatory myofibroblastic tumors
harbor multiple potentially actionable kinase fusions, Cancer Discov. 4 (2014)
889–895.
[3] A.T. Shaw, S.H. Ou, Y.J. Bang, et al., Crizotinib in ROS1-rearranged non-small-cell
lung cancer, N. Engl. J. Med. 371 (2014) 1963–1971.
[4] A.T. Shaw, E. Felip, T.M. Bauer, et al., Lorlatinib in non-small-cell lung cancer with
ALK or ROS1 rearrangement: an international, multicentre, open-label, single-arm
first-in-man phase 1 trial, Lancet Oncol. 18 (2017) 1590–1599.
[5] P. Schoffski, J. Sufliarsky, H. Gelderblom, et al., Crizotinib in patients with
advanced, inoperable inflammatory myofibroblastic tumours with and without
anaplastic lymphoma kinase gene alterations (European Organisation for Research
and Treatment of Cancer 90101 CREATE): a multicentre, single-drug, prospective,
non-randomised phase 2 trial, Lancet Respir. Med. 6 (2018) 431–441.
[6] S. Reinhart, Y. Trachsel, C. Fritz, et al., Inflammatory myofibroblastic tumor of the
bladder with FN1-ALK gene fusion: different response to ALK inhibition, Urology
146 (2020) 32–35.
[7] H.H. Wong, H. Bentley, V.R. Bulusu, et al., Lorlatinib for the treatment of
inflammatory myofibroblastic tumour with TPM4-ALK fusion following failure of
entrectinib, Anticancer Drugs 31 (2020) 1106–1110.
[8] M. Casanova, B. Brennan, R. Alaggio, et al., Inflammatory myofibroblastic tumor:
the experience of the European pediatric Soft tissue Sarcoma Study Group (EpSSG),
Eur. J. Cancer 127 (2020) 123–129.
[9] J.C. Chang, L. Zhang, A.E. Drilon, et al., Expanding the molecular characterization
of thoracic inflammatory myofibroblastic tumors beyond ALK gene
rearrangements, J. Thorac. Oncol. 14 (2019) 825–834.
[10] C.R. Antonescu, A.J. Suurmeijer, L. Zhang, et al., Molecular characterization of
inflammatory myofibroblastic tumors with frequent ALK and ROS1 gene fusions
and rare novel RET rearrangement, Am. J. Surg. Pathol. 39 (2015) 957–967.
K.M. Ingley et al.