Treatment Response In Solid Tumors Research Articles

Reproducing RECIST lesion selection via machine learning: Insights into intra and inter-radiologist variation

BackgroundThe Response Evaluation Criteria in Solid Tumors (RECIST) aims to provide a standardized approach to assess treatment response in solid tumors. However, discrepancies in the selection of measurable and target lesions among radiologists using these criteria pose a significant limitation to their reproducibility and accuracy. This study aimed to understand the factors contributing to this variability. MethodsMachine learning models were used to replicate, in parallel, the selection process of measurable and target lesions by two radiologists in a cohort of 40 patients from an internal pan-cancer dataset. The models were trained on lesion characteristics such as size, shape, texture, rank, and proximity to other lesions. Ablation experiments were conducted to evaluate the impact of lesion diameter, volume, and rank on the selection process. ResultsThe models successfully reproduced the selection of measurable lesions, relying primarily on size-related features. Similarly, the models reproduced target lesion selection, relying mostly on lesion rank. Beyond these features, the importance placed by different radiologists on different visual characteristics can vary, specifically when choosing target lesions. Worth noting that substantial variability was still observed between radiologists in both measurable and target lesion selection. ConclusionsDespite the successful replication of lesion selection, our results still revealed significant inter-radiologist disagreement. This underscores the necessity for more precise guidelines to standardize lesion selection processes and minimize reliance on individual interpretation and experience as a means to bridge existing ambiguities.

CtDNA: An emerging neoadjuvant biomarker in resectable solid tumors.

Christopher Abbosh and Charles Swanton discuss circulating tumor DNA as a potential biomarker for neoadjuvant treatment response in solid tumors.

116 PROGNOSTIC IMPACT OF THE SHORT-AXIS DIAMETER FOR THE NODAL ASSESSMENT BEFORE PREOPERATIVE CHEMOTHERAPY IN ESOPHAGEAL CANCER

Abstract Clinical N category (cN) is generally assessed by measuring the nodal diameter by CT before the initiation of primary treatment in esophageal squamous cell carcinoma (ESCC). The short-axis diameter is recommended for evaluating treatment response in solid tumors by RECIST. This study aimed to elucidate the prognostic implication of the maximum short-axis diameter of lymph node (cN-size) before preoperative chemotherapy for ESCC. Methods We enrolled a total of 152 patients who underwent preoperative cisplatin/5-fluorouracil therapy (CF) followed by esophagectomy from 2005 to 2011. There were 127 men and 25 women with a median age of 65 years (range: 47–79 years). Clinically metastatic node was defined as follows; the node with cN-size ≥10 mm or that with 5 mm ≤ cN-size <10 mm and contrast enhancement, round shape and/or central necrosis in CT before starting CF. The association between the maximum cN-size and the overall survival (OS) after surgery was statistically investigated. The median follow-up period was 87 months (range: 36–145 months). Results The number of patients with cN0 and cN1–3 was 60 and 92, respectively. Twenty-seven and 65 patients with cN1–3 were classified into cN-size <10 mm and cN-size ≥10 mm group, respectively. The 5-year OS rates in cN0, cN-size <10 mm and cN-size ≥10 mm groups were 70%, 51% and 45%, respectively (P = 0.006). Among Ut-Mt tumors, the OS in the cN-size <10 mm group was significantly worse than that in the cN0 group (5-year OS rate: 45% vs. 74%, P = 0.048). However, there were no significant differences in the OS between these two groups in Lt tumors (67% vs. 64%, P = 0.789). Conclusion The maximum short-axis diameter of lymph node before preoperative chemotherapy is significantly associated with OS in patients with ESCC. Lymph node with 5 mm ≤ cN-size <10 mm in the short axis should be treated as a metastatic node especially in Ut or Mt tumors, considering the poor prognosis.

Dynamics of circulating tumor DNA in patients with advanced solid tumors treated with cediranib and olaparib.

3035 Background: Circulating tumor DNA (ctDNA) has emerged as a potential biomarker to monitor treatment response in solid tumors. Our group previously showed that changes in ctDNA levels were predictive of radiographic response and survival in NSCLC patients receiving immunotherapy. Here we evaluated whether ctDNA dynamics could similarly be used to assess response in a PARP inhibitor-based therapy. Methods: A total of 122 patients with NSCLC, TNBC, PDAC or SCLC received cediranib (C) 30mg daily and Olaparib (O) 200mg twice daily in a phase II study NCI9881. Using a multigene NGS assay, ctDNA was measured longitudinally at baseline (T0), after 3 to 7 days of C monotherapy (T1), after 1 week of C+O combination (T2), after 4 weeks of C+O (T3), and every 4 weeks (T4+) thereafter. The first radiographic assessment was done after 8 weeks of C+O and every 8-12 weeks thereafter. CtDNA was quantified by determining the allele fraction of cancer-associated somatic mutations in plasma. We defined an early ctDNA response (e-ctDNA-R) as a >10% decrease in mutant allele fraction from T0 to T2, and an early ctDNA progression (e-ctDNA-P) as a > 10% increase; otherwise, it was stable ctDNA (e-ctDNA-S). Results: In total, 493 samples were analyzed from 94 patients, and 40 unique patients had both T0 and T2 ctDNA measurements, as well as corresponding radiographic assessments. These included 10 NSCLC, 17 TNBC, 3 SCLC and 10 PDAC. Of these patients, 4, 21, and 15 patients had PR, SD, and PD as best overall radiographic response respectively. Twenty-three (57.5%) patients had either e-ctDNA-R (17, 42.5%) or e-ctDNA-S (6, 15.0%), 18 (78.3%) of whom subsequently had either radiographic partial response (PR) or stable disease (SD). Seventeen (42.5%) patients had e-ctDNA-P, 10 (58.8%) of whom then had PD. A fair agreement was observed between e-ctDNA-R/S or e-ctDNA-P and radiographic PR/SD or PD with Cohen’s k 0.38 (70% agreement). The correlation between early ctDNA changes and PFS/OS are summarized in the table. All 25 patients with PR/SD eventually progressed. Of these, 17 (68%) had >10% increase in ctDNA from the nadir prior to disease progression (median 94.6%, 95%CI 38.2%-389.1%). The time between ctDNA progression and ctDNA nadir was significantly shorter (median 21 days, 95%CI 21-28) than the time between radiographic/clinical progression and initial PR/SD (median 107 days, 95%CI 56-204, P=0.0015). Conclusions: Longitudinal ctDNA measurements could enable early assessment of treatment response, resistance, and disease progression in patients treated with PARP inhibitor-based therapy. However, in this study, tumor responses and ctDNA changes were generally not as robust as have been observed with other classes of therapy.[Table: see text]

Predicting treatment response from longitudinal images using multi-task deep learning

Radiographic imaging is routinely used to evaluate treatment response in solid tumors. Current imaging response metrics do not reliably predict the underlying biological response. Here, we present a multi-task deep learning approach that allows simultaneous tumor segmentation and response prediction. We design two Siamese subnetworks that are joined at multiple layers, which enables integration of multi-scale feature representations and in-depth comparison of pre-treatment and post-treatment images. The network is trained using 2568 magnetic resonance imaging scans of 321 rectal cancer patients for predicting pathologic complete response after neoadjuvant chemoradiotherapy. In multi-institution validation, the imaging-based model achieves AUC of 0.95 (95% confidence interval: 0.91–0.98) and 0.92 (0.87–0.96) in two independent cohorts of 160 and 141 patients, respectively. When combined with blood-based tumor markers, the integrated model further improves prediction accuracy with AUC 0.97 (0.93–0.99). Our approach to capturing dynamic information in longitudinal images may be broadly used for screening, treatment response evaluation, disease monitoring, and surveillance.

Exploring fingerprint ctDNA as a stratifying factor in colorectal cancer.

e16101 Background: Colorectal cancer (CRC) ranks near the top in tumor-related deaths. Its standard-of-care is surgical resection plus adjuvant chemotherapy, which can extend patient’s live and disease-free survival for up to several years. However, choosing an appropriate chemotherapy agent is difficult because CRC is a heterogenous disease. It is believed that molecular stratification can help to improve treatment accuracy and this has been proven by the consensus molecular subtypes (CMS) classification. However, the CMS classification relies on expensive and complex gene-expression profiling, and more cumbersomely, is limited to early stage patients without prior chemotherapy, radiation and metastasis. This situation denies many CRC patients from the benefit of CMS classification. Circulating tumor DNA (ctDNA) is a minimally invasive way to monitor disease progression and treatment response in solid tumors but its clinical utility in CRC remains to be validated. We explore the potential of using patient’s fingerprint ctDNA (defined below) as a stratifying factor in CRC patients to assist clinical decision. Methods: WES is performed on tumor and matched blood samples from 149 CRC patients. Based on the WES result, 20-30 tumor specific genes are selected for each patient and form their ctDNA fingerprints. The patients are grouped according to their level of fingerprint ctDNA (high- vs. low-ctDNA). Results: The two groups of patients show significant difference in treatment responses and mutational profiles. The low-ctDNA group in general respond to treatment well and have good prognosis. The high-ctDNA group, in contrast, often experience relapse or recurrence. Interestingly, the low-ctDNA group is dominated by point and small indel mutations in the top mutated genes while the high-ctDNA group is dominated by gene copy variations (CNV). SMAD4 deficit and DCC amplification are well known CNV in CRC, but they only appear in the high-ctDNA group. Similarly, CNV of AGO2, ACTG1, MYC, and BRD3 are only associated with the high-ctDNA group but not the low-ctDNA group. Conclusions: Our result indicates a strong correlation of fingerprint ctDNA level to both tumor mutational profile and prognosis. This suggests fingerprint ctDNA level may be used as an effective stratifying factor in CRC. At the same time, possibly a common molecular cause is driving persistent ctDNA production in CRC patients, including a large portion of them who have received surgical and adjuvant chemotherapy. Further perspective study, however, is needed to test whether fingerprint ctDNA can be a predictive biomarker.

Enrichment and Characterization of the Tumor Immune and Non-immune Microenvironments in Established Subcutaneous Murine Tumors.

The tumor immune microenvironment (TIME) has recently been recognized as a critical mediator of treatment response in solid tumors, especially for immunotherapies. Recent clinical advances in immunotherapy highlight the need for reproducible methods to accurately and thoroughly characterize the tumor and its associated immune infiltrate. Tumor enzymatic digestion and flow cytometric analysis allow broad characterization of numerous immune cell subsets and phenotypes; however, depth of analysis is often limited by fluorophore restrictions on panel design and the need to acquire large tumor samples to observe rare immune populations of interest. Thus, we have developed an effective and high throughput method for separating and enriching the tumor immune infiltrate from the non-immune tumor components. The described tumor digestion and centrifugal density-based separation technique allows separate characterization of tumor and tumor immune infiltrate fractions and preserves cellular viability, and thus, provides a broad characterization of the tumor immunologic state. This method was used to characterize the extensive spatial immune heterogeneity in solid tumors, which further demonstrates the need for consistent whole tumor immunologic profiling techniques. Overall, this method provides an effective and adaptable technique for the immunologic characterization of subcutaneous solid murine tumors; as such, this tool can be used to better characterize the tumoral immunologic features and in the preclinical evaluation of novel immunotherapeutic strategies.

Abstract 4214: Assessment of early antimitotic treatment response in ovarian cancer using temporal diffusion spectroscopy

Abstract A quantitative MRI method, namely, temporal diffusion spectroscopy (TDS), has been used to measure tumor microstructural changes in response to a well-established antimitotic drug treatment (protein-bound paclitaxel) and evaluated as an early indicator of tumor responsiveness. Current radiological methods of monitoring treatment response in solid tumors rely on measuring changes in tumor size as measured by CT or MRI, and it is well recognized this approach suffers from major limitations. In recent years, tumor size measurements have been supplemented by a number of imaging techniques that aim for in vivo characterization and measurement of biologic processes at the cellular and molecular level, but to date none has proven robustly successful in practice. For example, diffusion MRI has shown a high sensitivity to tumor response, but a relatively lower specificity e.g. sometimes unable to distinguish real response of responder tumors from the progression of some non-responders. Temporal diffusion spectroscopy measures a series of apparent diffusion coefficients (ADC) over a wide range of effective diffusion times, corresponding to diffusion distances ranging from subcellular to cellular levels (∼ 3-20 μm). By fitting the measured ADC spectra to a simplified tissue model, parameters reflecting structural properties (such as cell size) in tissues can be extracted. Abraxane, one of the most successful microtubule-targeted chemotherapeutic drugs, has been reported to induce aberrant mitosis in tumor cells leads to mitotic arrest, the consequence of which is a cell size increase. In the current study, we implemented TDS in Abraxane/PBS-treated two types of human ovarian xenografts and cultured cells (OVCAR-8 as a responder to Abraxane, and NCI/ADR-RES as a resistant type). Abraxane-induced cell size increases were confirmed by flow cytometry and light microscopy in cell culture. The efficacy of Abraxane in treating tumors in vivo was examined histologically by calculating percentages of Caspase-3-activated area within Caspase-3 stained slides. Most excitingly, the fitted restricted size, one of the spectral parameters obtained from TDS, was able to quantify cell size increases which were not detected by conventional diffusion-MRI. All the MR results had a high degree of consistency with other flow, microscopy, and histological data. Moreover, with an appropriate analysis, the Abraxane-responsive tumors in vivo could be easily distinguished from all the other drug-vehicle-treated tumors and Abraxane-resistant tumors. In conclusion, temporal diffusion spectroscopy detects antimitotic-therapy-induced microstructural variations (notably, increases in cell sizes) in tumors in vivo which occur before changes in tumor volume or conventional diffusion MRI metrics. A combination of restricted size and zero-frequency diffusion coefficient can potentially be used as an early indicator of tumor responsiveness. Citation Format: Xiaoyu Jiang, Hua Li, Ping Zhao, Jingping Xie, Dineo Khabele, Junzhong Xu, John C. Gore. Assessment of early antimitotic treatment response in ovarian cancer using temporal diffusion spectroscopy. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4214.

The Role of Magnetic Resonance Imaging Biomarkers in Clinical Trials of Treatment Response in Cancer

Current standard-of-care radiological methods for assessing the response of solid tumors to treatment are based on measuring changes in lesion size in a single dimension using high-resolution x-ray computed tomography (CT) or magnetic resonance imaging (MRI). Even if size measurements are adapted to record true volume changes more accurately, the effects of therapeutic drugs on tumor size may not occur for several cycles of treatment. Furthermore, current and future generations of anticancer drugs will be designed to affect highly specific cancer characteristics, and their effects may not be immediately cytotoxic. More sensitive and specific measures are required that can report on tumor status and treatment response early in the course of therapy. Several MRI techniques have matured to the point where they can offer quantitative information on tissue status and greater insight into specific biophysical and physiological characteristics of tumors. Here we review and provide illustrative examples of two MRI methods that have already been incorporated into clinical trials of treatment response in solid tumors: diffusion imaging and dynamic contrast-enhanced MRI. We also discuss the limitations and future research directions required for these techniques to gain greater acceptance and to have their maximum impact.

WE‐C‐BRC‐04: Predictive Power of Various PET Measures for Treatment Response Assessment

Purpose: Assessment of treatment response in solid tumors is essential for disease management but typical response measures are limited and non‐robust. We compared the predictive power of several PET‐based measures in cancer patients receiving therapy. Method and Materials: Fourteen adult patients with advanced solid malignancies were treated with sunitinib malate, a molecular targeted agent. Using the cellular proliferation marker []FLT, whole‐body PET/CT scans were acquired pre‐, mid‐, and post‐treatment. Lesions were segmented on PET images and treatment response was assessed via percent change of the following measures: SUVmean, SUVmax, SUVpeak, SUVtotal, and PET defined tumor size: unidimensional, bidimensional, and volume. PET response assessment and clinical endpoint of therapy were used independently to classify patients into response categories: partial response (PR), stable disease (SD), or progressive disease (PD). Predictive power and robustness of PET measures were tested by varying response category cutoffs to generate ROC curves and Matthews correlation coefficients (MCC), a classification quality measure. Results: SUV measures demonstrated improved predictive power over PET defined size measures (MCCaverage: 0.27 vs. 0.21). Of all measures, SUVpeak was the most robust response predictor (MCCaverage: 0.30, MCCmax: 1.0), especially for prediction of PR. Of PET defined size measures, bidimensional size was the most robust response predictor. Changes in SUV measures pre‐ to mid‐treatment showed improved predictive power over pre‐ and mid‐ to post‐treatment changes (MCCaverage: 0.32, 0.27, and 0.22 respectively). Prediction of PR was vastly superior to that of SD and PD (MCCaverage: 0.44, 0.16, and 0.14 respectively). ROC analysis supported MCC results. Conclusion: Comparison of PET‐based measures revealed that SUV measures were more predictive of response than PET defined size measures. SUVpeak, a rarely used functional measure, demonstrated the most robust predictive power. Future treatment response studies should test and implement more robust measures, like SUVpeak, for improved response assessment.

ErbB1–4 expression in prostate cancer patients and its correlation to patients’ ethnicity and outcome

14615 Background: ErbB family is involved in both cancer progression and treatment response in solid tumors. Few inconclusive studies reported on ErbB over-expression in prostate cancer. We investigated ErbB1–4 expression in prostate cancer patients and its correlation to patients ethnicity and outcome. Methods: ErbB expression was evaluated by immunohistochemistry of prostate cancer specimen using polyclonal antibody (Santa Cruz, CA). The staining was recorded as negative (0/+1), moderately positive (+2) and highly positive (+3). Kattan nomogram was used to predict 5-yr progression-free probability, assuming that all patients received external beam radiation therapy (a total dose of 78 Gy) and hormonal manipulation. Origin was counted in all 43 patients: Ashkenazic patients were defined as those who immigrated from East/West Europe or North America and Sephardic patients - from Middle East and North Africa. Results: ErbB1 (+2/+3) was over-expressed in 12 and 7 patients for a total of 19/43 (44%). ErbB2 over-expression (+2/+3) was not found in all patients. ErbB3 over-expression of +2 was seen in 2 patients and none had +3 (2/43, 5%). ErbB4 over-expression (+2/+3) was seen in 5 and 11 patients for a total of 16/43 (37%). 22 patients were Ashkenazic and 21 - Sephardic. ErbB1 over-expression in Ashkenazic and Sephardic groups was 9/22 (41%) and 10/21 (48%). ErbB4 over-expression in the two groups was 7/22 (32%) and 9/21 (43%). Kattan score of <80 was seen in 20/43 and <60 in 7/43 patients. ErbB1 over-expression was noted in 11/20 and in 4/7 patients. ErbB4 over-expression was seen in 7/20 and in 4/7 patients. In both ErbB1 and ErbB4 over-expression and Kattan nomogram of <80 and <60 the Sephardic ethnicity dominated-7/11 (64%), 3/4 (75%), 5/7 (71%) and 3/4 (75%). Conclusions: ErbB1 and ErbB4 over-expression is presented in 43% and 37% patients while ErbB3 was over-expressed in 5%; no over-expression of ErbB2 was observed. Ashkenazic and Sephardic ethnicity were evenly distributed in the over-expressed ErbB1 and ErbB4 patients. However, a tendency to a worse prognosis, based on Kattan nomogram, was seen in over-expressed ErbB1 and ErbB4 patients from Sephardic ethnicity. Further studies on ethnicity and ErbB prevalence and prognosis are warranted. No significant financial relationships to disclose.