Resistance to osimertinib in advanced EGFR-mutated NSCLC: a prospective study of molecular genotyping on tissue and liquid biopsies

BACKGROUND

Resistance to osimertinib in advanced EGFR-mutated non-small cell lung cancer (NSCLC) constitutes a significant challenge for clinicians either in terms of molecular diagnosis and subsequent therapeutic implications.

METHODS

This is a prospective single-centre study with the primary objective of characterising resistance mechanisms to osimertinib in advanced EGFR-mutated NSCLC patients treated both in first- and in second-line. Next-Generation Sequencing analysis was conducted on paired tissue biopsies and plasma samples. A concordance analysis between tissue and plasma was performed.

RESULTS

Sixty-five advanced EGFR-mutated NSCLC patients treated with osimertinib in first- (n = 56) or in second-line (n = 9) were included. We managed to perform tissue and liquid biopsies in 65.5% and 89.7% of patients who experienced osimertinib progression, respectively. Acquired resistance mechanisms were identified in 80% of 25 patients with post-progression samples, with MET amplification (n = 8), EGFR C797S (n = 3), and SCLC transformation (n = 2) the most frequently identified. The mean concordance rates between tissue and plasma for the EGFR activating mutation and for the molecular resistance mechanisms were 87.5% and 22.7%, respectively.

CONCLUSIONS

Resistance to osimertinib demonstrated to be highly heterogeneous, with MET amplification the main mechanism. Plasma genotyping is a relevant complementary tool which might integrate tissue analysis for the study of resistance mechanisms.

PD-L1 overexpression induces STAT signaling and promotes the secretion of pro-angiogenic cytokines in non-small cell lung cancer (NSCLC)

BACKGROUND

Monoclonal antibodies (ICI) targeting the immune checkpoint PD-1/PD-L1 alone or in combination with chemotherapy have demonstrated relevant benefits and established new standards of care in first-line treatment for advanced non-oncogene addicted non-small cell lung cancer (NSCLC). However, a relevant percentage of NSCLC patients, even with high PD-L1 expression, did not respond to ICI, highlighting the presence of intracellular resistance mechanisms that could be dependent on high PD-L1 levels. The intracellular signaling induced by PD-L1 in tumor cells and their correlation with angiogenic signaling pathways are not yet fully elucidated.

METHODS

The intrinsic role of PD-L1 was initially checked in two PD-L1 overexpressing NSCLC cells by transcriptome profile and kinase array. The correlation of PD-L1 with VEGF, PECAM-1, and angiogenesis was evaluated in a cohort of advanced NSCLC patients. The secreted cytokines involved in tumor angiogenesis were assessed by Luminex assay and their effect on Huvec migration by a non-contact co-culture system.

RESULTS

PD-L1 overexpressing cells modulated pathways involved in tumor inflammation and JAK-STAT signaling. In NSCLC patients, PD-L1 expression was correlated with high tumor intra-vasculature. When challenged with PBMC, PD-L1 overexpressing cells produced higher levels of pro-angiogenic factors compared to parental cells, as a consequence of STAT signaling activation. This increased production of cytokines involved in tumor angiogenesis largely stimulated Huvec migration. Finally, the addition of the anti-antiangiogenic agent nintedanib significantly reduced the spread of Huvec cells when exposed to high levels of pro-angiogenic factors.

CONCLUSIONS

In this study, we reported that high PD-L1 modulates STAT signaling in the presence of PBMC and induces pro-angiogenic factor secretion. This could enforce the role of PD-L1 as a crucial regulator of the tumor microenvironment stimulating tumor progression, both as an inhibitor of T-cell activity and as a promoter of tumor angiogenesis.

P01.14.B Visuospatial selective attention deficit after resection of right hemisphere gliomas: a multivariate lesion symptom mapping and diffusion tractography study

Background

The ability to select relevant information to current behavioural goals despite concomitant distractors (i.e. selective attention) is crucial for daily life. Despite the implication of fronto-parietal network in sustaining visuospatial selective attention has been considered crucial, the role of subcortical structures is still unclear. Aim of this study was to evaluate the impact of neurosurgical removal of right gliomas on visuospatial selective attention performances.

Material and Methods

The visuospatial selective attention domain was assessed pre- and postoperatively (1 month follow-up) in 73 patients operated for right hemisphere gliomas. Support-vector-regression lesion-symptom mapping and disconnectome analysis (Lesion Quantification Toolkit, HCP 1065 data) were used to evaluate the relationship between the resection cavities and the postoperative performance on a standardized selective attention test. To test and refine the atlas-based results, in a subset of 17 patients with preoperative diffusion tractography, the postoperative outcome was correlated with the percentage of disconnection for different white matter tracts.

Results

Results showed that resection of a cluster enclosing the SMA-complex and the surrounding white matter was significantly associated with a decline in visuospatial selective attention performances (p<.05). Disconnectome analysis at population level revealed that the deficit was significantly correlated with the resection of superior cortico-striatal fibres, superior thalamic radiations, corticopontine projections, callosal connections and fibres of the frontal aslant tract. Within all the cortico-subcortical terminations of these tracts, disconnection of streamlines of the SMA-complex was the most associated with selective attention deficits. Tractography in single patients confirmed the population level results and suggested that ventro-lateral resections in case of inferior frontal or insular tumours significantly disconnecting the frontal aslant tract but sparing superior projection fibres were not associated with any postoperative disturbances.

Conclusion

Results show converging evidence to support a critical role for the SMA/preSMA and its projections in maintaining efficiency of attentive control. The present results reveal the importance of a fronto-medial descending connectivity in mediating the voluntary control of visuospatial selective attention. From a surgical standpoint, to avoid postoperative deficits, it seems crucial to preserve SMA projection pathways, while the disconnection of the frontal aslant tract is feasible without any postoperative decline.

P01.19.A The temporo-parieto-occipital junction in the non-dominant hemisphere as a challenging hub for glioma surgery: functional and oncological outcomes

Background

The temporo-parieto-occipital junction (TPOJ) in the non-dominant hemisphere is a complex region intersected by multiple white matter bundles (Optic radiations, Inferior-fronto-occipital tract, inferior-longitudinalis-tract, superior-longitudinalis-tract, vertical-occipital-tract, medial-dorsal-longitudinal-tract, parietal-aslant-tract), subserving several high-level neurological functions such as spatial recognition, visual agnosia, visual field, attention and working memory. Few data are available on the optimum brain mapping strategy to be adopted for resecting tumors involving this area. Ideally it should allow surgeon to effectively identify several tracts and safely preserve all functions.

Material and Methods

We developed a proper protocol to approach this complex area integrating, in awake condition, intraoperative visual test (iVT) to mapping visual field, semantic association test (SRAT) to prevent visual agnosia and preserve working memory, hand manipulation task (hMT) to preserve spatial abilities. We reviewed its efficacy in a series of 38 patients with tumors involving non-dominant TPOJ, looking to functional (neurologic-neuropsychological) and oncological (EOR) outcomes. we perform a lesion symptom map and a disconnectome analysis to evaluate what region predict a decline in neuropsychological function and which tract of white brain matter are correlated in the decline of performance.

Results

Feasibility was high and all patients were able to perform and complete the protocol, which lasted, on the average11 min cortically and 25 min subcortically. Specificity was >95%. Immediate post-operative deficits were documented in 87.4% of patients, permanent in 3.9% (visual field, visual spatial abilities). Attentive and emotional domains were those mostly affected in the neuropsychological evaluation. Lesion symptom mapping and disconnectomic analysis showed that the postoperative decrease in neuropsychological performance was associated with resection of a cluster of voxels corresponding to the anterior portion of the temporo-parietal junction. The white matter tracts mainly involved were the anterior and posterior segment of the middle longitudinal fasciculus (aMdLF and pMdLF) and the parietal aslant tract (PAT).

Conclusion

Effective and safe resection of tumors involving non-dominant TPOJ is feasible. The adoption of a specific brain mapping protocol in awake setting is recommended to achieve a full functional preservation and extend tumor resection. Further analysis should be performed to assess the role of this subcortical tracts.

Soluble PD-L1 and Circulating CD8+PD-1+ and NK Cells Enclose a Prognostic and Predictive Immune Effector Score in Immunotherapy Treated NSCLC patients

INTRODUCTION

Upfront criteria to foresee immune checkpoint inhibitors (ICIs) efficacy are far from being identified. Thus, we integrated blood descriptors of pro-inflammatory/immunosuppressive or effective anti-tumor response to non-invasively define predictive immune profiles in ICI-treated advanced non-small cell lung cancer (NSCLC).

METHODS

Peripheral blood (PB) was prospectively collected at baseline from 109 consecutive NSCLC patients undergoing ICIs as first or more line treatment. Soluble PD-L1 (sPD-L1) (immunoassay), CD8+PD-1+ and NK (FACS) cells were assessed and interlaced to generate an Immune effector Score (I_effS). Lung Immune Prognostic Index (LIPI) was computed by LDH levels and derived Neutrophil-to-Lymphocyte Ratio (dNLR). All these parameters were correlated with survival outcome and treatment response.

RESULTS

High sPD-L1 and low CD8+PD-1+ and NK number had negative impact on PFS (P < 0.001), OS (P < 0.01) and ICI-response (P < 0.05). Thus, sPD-L1^high, CD8+PD-1+^low and NK^low were considered as risk factors encompassing I_effS, whose prognostic power outperformed that of individual features and slightly exceeded that of LIPI. Accordingly, the absence of these risk factors portrayed a favorable I_effS characterizing patients with significantly (P < 0.001) prolonged PFS (median NR vs 2.3 months) and OS (median NR vs 4.1) and greater benefit from ICIs (P < 0.01). We then combined each risk parameter composing I_effS and LIPI (LDH^high, dNLR^high), thus defining three distinct prognostic classes. A remarkable impact of I_effS-LIPI integration was documented on survival outcome (PFS, HR = 4.61; 95%CI = 2.32-9.18; P < 0.001; OS, HR=4.03; 95%CI=1.91-8.67; P < 0.001) and ICI-response (AUC=0.90, 95%CI=0.81-0.97, P < 0.001).

CONCLUSION

Composite risk models based on blood parameters featuring the tumor-host interaction might provide accurate prognostic scores able to predict ICI benefit in NSCLC patients.

Psychiatric illness and subsequent traumatic brain injury: a case control study

OBJECTIVE

To determine whether psychiatric illness is a risk factor for subsequent traumatic brain injury (TBI).

METHODS

Case control study in a large staff model health maintenance organisation in western Washington State. Patients with TBI, determined by International classification of diseases, 9th revision, clinical modification (ICD-9-CM) diagnoses, were 1440 health plan members who had TBI diagnosed in 1993 and who had been enrolled in the previous year, during which no TBI was ascertained. Three health plan members were randomly selected as control subjects, matched by age, sex, and reference date. Psychiatric illness in the year before the TBI reference date was determined by using computerised records of ICD-9-CM diagnoses, psychiatric medication prescriptions, and utilisation of a psychiatric service.

RESULTS

For those with a psychiatric diagnosis in the year before the reference date, the adjusted relative risk for TBI was 1.7 (95% confidence interval (CI) 1.4 to 2.0) compared with those without a psychiatric diagnosis. Patients who had filled a psychiatric medication prescription had an adjusted relative risk for TBI of 1.6 (95% CI 1.2 to 2.1) compared with those who had not filled a psychiatric medication prescription. Patients who had utilised psychiatric services had an adjusted relative risk for TBI of 1.3 (95% CI 1.0 to 1.6) compared with those who had not utilised psychiatric services. The adjusted relative risk for TBI for patients with psychiatric illness determined by any of the three psychiatric indicators was 1.6 (95% CI 1.4 to 1.9) compared with those without any psychiatric indicator.

CONCLUSION

Psychiatric illness appears to be associated with an increased risk for TBI.