Artificial intelligence for predictive biomarker discovery in immuno-oncology: a systematic review

BACKGROUND

The widespread use of immune checkpoint inhibitors (ICIs) has revolutionised treatment of multiple cancer types. However, selecting patients who may benefit from ICI remains challenging. Artificial intelligence (AI) approaches allow exploitation of high-dimension oncological data in research and development of precision immuno-oncology.

MATERIALS AND METHODS

We conducted a systematic literature review of peer-reviewed original articles studying the ICI efficacy prediction in cancer patients across five data modalities: genomics (including genomics, transcriptomics, and epigenomics), radiomics, digital pathology (pathomics), and real-world and multimodality data.

RESULTS

A total of 90 studies were included in this systematic review, with 80% published in 2021-2022. Among them, 37 studies included genomic, 20 radiomic, 8 pathomic, 20 real-world, and 5 multimodal data. Standard machine learning (ML) methods were used in 72% of studies, deep learning (DL) methods in 22%, and both in 6%. The most frequently studied cancer type was non-small-cell lung cancer (36%), followed by melanoma (16%), while 25% included pan-cancer studies. No prospective study design incorporated AI-based methodologies from the outset; rather, all implemented AI as a post hoc analysis. Novel biomarkers for ICI in radiomics and pathomics were identified using AI approaches, and molecular biomarkers have expanded past genomics into transcriptomics and epigenomics. Finally, complex algorithms and new types of AI-based markers, such as meta-biomarkers, are emerging by integrating multimodal/multi-omics data.

CONCLUSION

AI-based methods have expanded the horizon for biomarker discovery, demonstrating the power of integrating multimodal data from existing datasets to discover new meta-biomarkers. While most of the included studies showed promise for AI-based prediction of benefit from immunotherapy, none provided high-level evidence for immediate practice change. A priori planned prospective trial designs are needed to cover all lifecycle steps of these software biomarkers, from development and validation to integration into clinical practice.

Stereo-EEG tailored resection in a child with presumed perinatal post-stroke epilepsy: The complex organization of epileptogenic zone

Introduction

Only a few studies have focused on tailored resection in post-stroke epilepsy, in which hemispherectomy and hemispherotomy are the most recognized treatments.

Case description

We describe the case of a patient with drug-resistant, presumed perinatal, post-stroke epilepsy and moderate right hemiparesis. The seizures were stereotyped, both spontaneous and induced by sudden noises and somatosensory stimuli. Considering the discordant anatomic-electro-clinical data - left perisylvian malacic lesion with electrical onset over the left mesial fronto-central leads - and the patient's functional preservation, SEEG was performed. SEEG revealed sub-continuous abnormalities in the perilesional regions. Several seizures were recorded, with onset over the premotor area, rapidly involving the motor and insular-opercular regions. We decided for a combined surgical approach, SEEG-guided radiofrequency thermocoagulation, on the fronto-mesial structure but also on the central operculum, followed by resective surgery including only the fronto-mesial structures.

Discussion and conclusion

The SEEG allowed to localize the epileptogenic zone far away from the anatomical lesion but connected to part of it. A combined surgical approach tailored on SEEG results allowed a good outcome (Engel Ib) without additional deficits.

STYLE (NCT03449173) a phase II Trial of Sunitinib in patients with type B3 Thymoma or Thymic Carcinoma in second and further lines

BACKGROUND

Thymic malignancies are rare tumors with few therapeutic options. The STYLE trial was aimed to evaluate activity and safety of sunitinib in advanced/recurrent type B3 thymoma (T) and thymic carcinoma (TC).

METHODS

In this multicenter Simon 2 stages phase II trial, patients with pretreated T/TC were enrolled in two cohorts and assessed separately. Sunitinib was administered 50mg daily for 4-weeks, followed by a 2-week rest period (schedule 4/2), until disease progression or unacceptable toxicity. The primary endpoint was objectives response rate (ORR). Progression free survival (PFS), overall survival (OS), disease control rate (DCR) and safety were secondary endpoints.

RESULTS

From 03/2017 to 01/2022, 12 T and 32 TC patients were enrolled. At stage 1 ORR was 0% (90%CI 0.0-22.1) in T and 16.7% (90%CI 3.1-43.8) in TC, so T cohort was closed. At stage 2, the primary endpoint was met for TC with ORR of 21.7% (90%CI 9.0%-40.4%). In the ITT analysis DCR was 91.7% (95%CI 61.5%-99.8%) in Ts and 89.3% (95%CI 71.8%-97.7%) in TCs. mPFS was 7.7 months (95%CI 2.4-45.5) in Ts, 8.8 (95%CI 5.3-11.1) in TCs; mOS was 47.9 months (95%CI 4.5-not reached) in Ts, and 27.8 (95%CI 13.2-53.2) in TCs. Adverse events (AEs) occurred in 91.7% T and 93.5% TCs. Grade ≥3 treatment-related AEs were reported in 25.0% T and 51.6% TC.

CONCLUSIONS

This trial confirms the activity of sunitinib in TC patients, supporting its use as second line treatment.

PEOPLE (NCT03447678), a first-line phase II pembrolizumab trial, in negative and low PD-L1 advanced NSCLC: clinical outcomes and association with circulating immune biomarkers

BACKGROUND

The PEOPLE trial aimed to identify new immune biomarkers in negative and low programmed death-ligand 1 (PD-L1) (0%-49%) advanced non-small-cell lung cancer (aNSCLC) patients treated with first-line pembrolizumab. Here we report the main outcomes and the circulating immune biomarkers analysis.

PATIENTS AND METHODS

The primary endpoint of this phase II trial was the identification of immune biomarkers associated with progression-free survival (PFS). Overall survival (OS), objective response rate (ORR), disease control rate (DCR), duration of response (DoR) and safety were secondary endpoints. Absolute cell counts for 36 subsets belonging to innate and adaptive immunity were determined by multiparametric flow cytometry in peripheral blood at baseline and at first radiologic evaluation. An orthoblique principal components-based clustering approach and multivariable Cox regression model adjusted for clinical variables were used to analyze immune variables and their correlation with clinical endpoints.

RESULTS

From May 2018 to October 2020, 65 patients were enrolled. After a median follow-up of 26.4 months, the median PFS was 2.9 months [95% confidence interval (CI) 1.8-5.6 months] and median OS was 12.1 months (95% CI 8.7-17.1 months). The ORR was 21.5%, DCR was 47.7% and median DoR was 14.5 months (95% CI 6.4-24.9 months). Drug-related grade 3-4 adverse events were 9.2%. Higher T cell and natural killer (NK) cell count at baseline and at the first radiologic evaluation were associated with improved PFS, DCR and OS. On the contrary, higher myeloid cell count at baseline or at the first radiologic evaluation was significantly associated with worse OS and DCR.

CONCLUSIONS

Circulating immune biomarkers can contribute to predict outcomes in negative and low PD-L1 aNSCLC patients treated with first-line single-agent pembrolizumab.

Role of impaired iron transport on exercise performance in heart failure patients

Aims

Impaired iron transport (IIT) occurs frequently in heart failure (HF) patients, even in the absence of anaemia and it is associated with a poor quality of life and prognosis. The impact of IIT on exercise capacity, as assessed by the cardiopulmonary exercise test (CPET), in HF is at present unknown. The aim of this article is to evaluate in HF patients the impact on exercise performance of IIT, defined as transferrin saturation (TSAT) <20%.

Methods and results

We collected data of 676 patients hospitalised for HF. All underwent laboratory analysis, cardiac ultrasound, and CPET. Patients were grouped by the presence/absence of IIT and anaemia (haemoglobin <13 and <12 g/dL in male and female, respectively): Group 1 (G1) no anaemia, no IIT; Group 2 (G2) anaemia, no IIT; Group 3 (G3) no anaemia, IIT; Group 4 (G4) anaemia and IIT. Peak oxygen uptake (peakVO2) reduced from G1 to G3 and from G2 to G4 (G1: 1266±497 mL/min, G2: 1011±385 mL/min, G3: 1041±395 mL/min, G4: 833±241 mL/min), whereas the ventilation to carbon dioxide relationship slope (VE/VCO2 slope) increased (G1: 31.8±7.5, G2: 34.5±7.4, G3: 36.1±10.2, G4: 37.5±8.4). At multivariate regression analysis, peakVO2 independent predictors were anaemia, brain natriuretic peptide (BNP), and left ventricular ejection fraction, whereas VE/VCO2 slope independent predictors were IIT and BNP.

Conclusions

In HF IIT is associated with exercise performance impairment independently from anaemia, and it is a predictor of elevated VE/VCO2 slope, a pivotal index of HF prognosis.

Funding Acknowledgement

Type of funding sources: Private hospital(s). Main funding source(s): Centro cardiologico Monzino, Italy

Evaluation of COVID-19 impact on DELAYing diagnostic-therapeutic pathways of lung cancer patients in Italy (COVID-DELAY study): fewer cases and higher stages from a real-world scenario

Introduction

COVID-19 has disrupted the global health care system since March 2020. Lung cancer (LC) patients (pts) represent a vulnerable population highly affected by the pandemic. This multicenter Italian study aimed to evaluate whether the COVID-19 outbreak had an impact on access to cancer diagnosis and treatment of LC pts compared with pre-pandemic time.

Methods

Consecutive newly diagnosed LC pts referred to 25 Italian Oncology Departments between March and December 2020 were included. Access rate and temporal intervals between date of symptoms onset and diagnostic and therapeutic services were compared with the same period in 2019. Differences between the 2 years were analyzed using the chi-square test for categorical variables and the Mann–Whitney U test for continuous variables.

Results

A slight reduction (−6.9%) in newly diagnosed LC cases was observed in 2020 compared with 2019 (1523 versus 1637, P = 0.09). Newly diagnosed LC pts in 2020 were more likely to be diagnosed with stage IV disease (P < 0.01) and to be current smokers (someone who has smoked more than 100 cigarettes, including hand-rolled cigarettes, cigars, cigarillos, in their lifetime and has smoked in the last 28 days) (P < 0.01). The drop in terms of new diagnoses was greater in the lockdown period (percentage drop −12% versus −3.2%) compared with the other months included. More LC pts were referred to a low/medium volume hospital in 2020 compared with 2019 (P = 0.01). No differences emerged in terms of interval between symptoms onset and radiological diagnosis (P = 0.94), symptoms onset and cytohistological diagnosis (P = 0.92), symptoms onset and treatment start (P = 0.40), and treatment start and first radiological revaluation (P = 0.36).

Conclusions

Our study pointed out a reduction of new diagnoses with a shift towards higher stage at diagnosis for LC pts in 2020. Despite this, the measures adopted by Italian Oncology Departments ensured the maintenance of the diagnostic-therapeutic pathways of LC pts.

•

The COVID-19 outbreak had an impact on access to lung cancer (LC) diagnosis and treatment.

•

A slight reduction (−6.9%) in newly diagnosed LC cases was observed in 2020 compared with 2019.

•

Newly diagnosed LC pts in 2020 were more likely to be diagnosed with stage IV disease.

•

The Italian Oncology Departments ensured the maintenance of the diagnostic-therapeutic pathways of LC pts.

•

A reverse migration from high-volume to low-volume cancer centers was noted during the pandemic.

Focal lesions induce large-scale percolation of sleep-like intracerebral activity in awake humans

Focal cortical lesions are known to result in large-scale functional alterations involving distant areas; however, little is known about the electrophysiological mechanisms underlying these network effects. Here, we addressed this issue by analysing the short and long distance intracranial effects of controlled structural lesions in humans. The changes in Stereo-Electroencephalographic (SEEG) activity after Radiofrequency-Thermocoagulation (RFTC) recorded in 21 epileptic subjects were assessed with respect to baseline resting wakefulness and sleep activity. In addition, Cortico-Cortical Evoked Potentials (CCEPs) recorded before the lesion were employed to interpret these changes with respect to individual long-range connectivity patterns. We found that small structural ablations lead to the generation and large-scale propagation of sleep-like slow waves within the awake brain. These slow waves match those recorded in the same subjects during sleep, are prevalent in perilesional areas, but can percolate up to distances of 60 mm through specific long-range connections, as predicted by CCEPs. Given the known impact of slow waves on information processing and cortical plasticity, demonstrating their intrusion and percolation within the awake brain add key elements to our understanding of network dysfunction after cortical injuries.

The lung immuno-oncology prognostic score (LIPS-3): a prognostic classification of patients receiving first-line pembrolizumab for PD-L1 ≥ 50% advanced non-small-cell lung cancer

Background

To stratify the prognosis of patients with programmed cell death-ligand 1 (PD-L1) ≥ 50% advanced non-small-cell lung cancer (aNSCLC) treated with first-line immunotherapy.

Methods

Baseline clinical prognostic factors, the neutrophil-to-lymphocyte ratio (NLR), PD-L1 tumour cell expression level, lactate dehydrogenase (LDH) and their combination were investigated by a retrospective analysis of 784 patients divided between statistically powered training (n = 201) and validation (n = 583) cohorts. Cut-offs were explored by receiver operating characteristic (ROC) curves and a risk model built with validated independent factors by multivariate analysis.

Results

NLR < 4 was a significant prognostic factor in both cohorts (P < 0.001). It represented 53% of patients in the validation cohort, with 1-year overall survival (OS) of 76.6% versus 44.8% with NLR > 4, in the validation series. The addition of PD-L1 ≥ 80% (21% of patients) or LDH < 252 U/l (25%) to NLR < 4 did not result in better 1-year OS (of 72.6% and 74.1%, respectively, in the validation cohort). Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 2 [P < 0.001, hazard ratio (HR) 2.04], pretreatment steroids (P < 0.001, HR 1.67) and NLR < 4 (P < 0.001, HR 2.29) resulted in independent prognostic factors. A risk model with these three factors, namely, the lung immuno-oncology prognostic score (LIPS)-3, accurately stratified three OS risk-validated categories of patients: favourable (0 risk factors, 40%, 1-year OS of 78.2% in the whole series), intermediate (1 or 2 risk factors, 54%, 1-year OS 53.8%) and poor (>2 risk factors, 5%, 1-year OS 10.7%) prognosis.

Conclusions

We advocate the use of LIPS-3 as an easy-to-assess and inexpensive adjuvant prognostic tool for patients with PD-L1 ≥ 50% aNSCLC.

•

Immunotherapy/chemoimmunotherapy combinations are currently not superior to immunotherapy alone for high PD-L1 aNSCLC.

•

NLR with a cut-off of 4 was validated as an independent prognostic factor for immunotherapy in high PD-L1 aNSCLC.

•

The addition of either PD-L1 ≥ 80% or LDH < 252 U/l to NLR < 4 did not result in better prognostic stratification.

•

The LIPS-3 is a validated 3-class prognostic classification based on the NLR, ECOG PS and pretreatment steroids.

•

The LIPS-3 is a routinely assessable adjuvant prognostic tool for high PD-L1 aNSCLC patients.

Single-arm, open label prospective trial to assess prediction of the role of ERCC1/XPF complex in the response of advanced NSCLC patients to platinum-based chemotherapy

Background

Platinum-based therapy, combined or not with immune checkpoint inhibitors, represents a front-line choice for patients with non-small-cell lung cancer (NSCLC). Despite the improved outcomes in the last years for this malignancy, only a sub-group of patients have long-term benefit. Excision repair cross-complementation group 1 (ERCC1) has been considered a potential biomarker to predict the outcome of platinum-based chemotherapy in NSCLC. However, the ERCC1 gene is transcribed in four splice variants where the isoform 202 was described as the only one active and able to complex Xeroderma pigmentosum group F-complementing protein (XPF). Here, we prospectively investigated if the active form of ERCC1, as assessed by the ERCC1/XPF complex (ERCC1/XPF), could predict the sensitivity to platinum compounds.

Patients and methods

Prospectively enrolled, patients with advanced NSCLC treated with a first-line regimen containing platinum were centrally evaluated for ERCC1/XPF by a proximity ligation assay. Overall survival (OS), progression-free survival (PFS) and objective response rate (ORR) were analyzed.

Results

The absence of the ERCC1/XPF in the tumor suggested a trend of worst outcomes in terms of both OS [hazard ratio (HR) 1.41, 95% confidence interval (CI) 0.67-2.94, P = 0.373] and PFS (HR 1.61, 95% CI 0.88-3.03, P = 0.123). ORR was marginally influenced in ERCC1/XPF-negative and -positive groups [odds ratio (stable disease + progressive disease versus complete response + partial response) 0.87, 95% CI 0.25-3.07, P = 0.832].

Conclusion

The lack of ERCC1/XPF complex in NSCLC tumor cells might delineate a group of patients with poor outcomes when treated with platinum compounds. ERCC1/XPF absence might well identify patients for whom a different therapeutic approach could be necessary.

•

This is the first study investigating the ERCC1/XPF complex as a platinum-based therapy response biomarker in NSCLC.

•

The lack of ERCC1/XPF complex might delineate a group of patients with poor outcomes when treated with platinum compounds.

•

ERCC1/XPF absence might identify tumors for whom a different therapeutic approach than platinum compounds could be necessary.

Determination of antioxidant capacity and flavonoid composition of onion (Allium cepa L.) landrace 'Krishnapuram' bulb using HPLC-ESI-ITMS

This study reports for the first-time antioxidant activity and flavonoid composition of KP onion landrace which is useful for future breeding programs and to obtain geographical indication (GI) tag for the benefit of farmers. The present study was aimed to determine antioxidant capacity and flavonoid composition of bulbs of red onion (Allium cepa L.) landrace 'Krishnapuram' (KP) from India using high-performance liquid chromatography (HPLC)-Electrospray Ionization (ESI)-multistage Ion Trap Mass Spectrometry (ITMS). The antioxidant activity was assayed by Ferric Reducing Antioxidant Power (FRAP) and hypochlorous acid (HClO)-induced oxidative damage in human erythrocytes. The total phenolic (TPC) contents in KP onion bulb extract (with 80% methanol) found to be 1.10 ± 0.2 mg GAE/g FW and 38.88 ± 1.0 lM QE/g. The FRAP activity measured for the bulb extract was 13.20 ± 0.1 μM QE/g. KP onion bulb extracts protected red blood cells (RBC) effectively (23%) against the oxidative damage induced by HClO. HPLC-ESI-ITMS analysis showed the presence of eight flavonols and five anthocyanins. Quercetin 3,4' -O-diglucoside (384.71 ± 0.49 mg/kg FW) and cyanidin 3-(6″-malonylglucoside) (20.95 ± 0.60 mg/kg FW) were detected as major flavonol and anthocyanin, respectively. The study suggests that KP onion has a considerable antioxidant activity due to the presence of high TPC. Moreover, quercetin glucosides are found to be more abundant than quercetin. The differences in quercetin glycosides content among different red onions could be useful for breeding programmes in the future.

Choosing wisely first line immunotherapy in non-small cell lung cancer (NSCLC): what to add and what to leave out

Immunotherapy has dramatically changed the therapeutic scenario in treatment naïve advanced non-small cell lung cancer (NSCLC). While single agent pembrolizumab has become the standard therapy in patients with PD-L1 expression on tumor cells ≥ 50%, the combination of pembrolizumab or atezolizumab and platinum-based chemotherapy has emerged as an effective first line treatment regardless of PD-L1 expression both in squamous and non-squamous NSCLC without oncogenic drivers. Furthermore, double immune checkpoint inhibition has shown promising results in treatment naïve patients with high tumor mutational burden (TMB). Of note, the presence of both negative PD-L1 expression and low TMB may identify a subgroup of patients who has little benefit from immunotherapy combinations and for whom the best treatment option may still be platinum-based chemotherapy. To date, first-line single agent immune checkpoint blockade has demonstrated limited activity in EGFR mutated NSCLC and the combination of immunotherapy and targeted agents has raised safety concerns in both EGFR and ALK positive NSCLC patients. Finally, in EGFR mutated or ALK rearranged NSCLC, atezolizumab in combination with platinum-based chemotherapy and bevacizumab is emerging as a potential treatment option upon progression to first line tyrosine kinase inhibitors.