1. IntroductionOncology has witnessed unprecedented growth in the understanding of aberrant molecular pathways implicated in tumourgenesis [1]. Advances in high-throughput genome sequencing technology, including next-generation sequencing (NGS), have enabled scientists to identify molecular characteristics associated with disease surveillance, treatment, and clinical outcomes [1,2]. Genomic testing in cancer detects significant somatic mutations to predict disease trajectory more accurately and to improve clinical outcomes through the development of guided personalized therapies [3,4]. Greater use of such testing to drive treatment decisions regarding solid tumours has enabled the shift in cancer diagnosis and therapy from empirical to precision oncology, intending to deliver the right cancer treatment to the right patient at the right dose and at the right time [3,4].Precision oncology testing is now a routine part of the initial assessment of many tumour types. For instance, in lung cancer, detecting EGFR mutations and ALK or ROS1 fusions helps to determine if a tyrosine kinase inhibitor (TKI) should be used instead of cytotoxic chemotherapy [5]. Additionally, PD-L1 expression (tumour proportion score of ≥50%) serves as a biomarker of the response to immunotherapy [6,7,8] in the lung, breast and many other cancer types. Similarly, among patients with metastatic colorectal cancer (mCRC), a RAS mutation is used as a predictive biomarker to select patients who will not benefit from epidermal growth factor receptor (EGFR) antibody therapy [9,10,11]. Such treatments driven by genomic testing have produced a marked increases in survival by altering the treatment decisions in approximately 50% of patients [6,7,8]. Furthermore, studies have shown that the successful use of these approaches can be safe and is independently associated with improvement across many outcome variables [11,12,13,14]. Despite the potential benefits of genomic testing, there is a growing clinical challenge to provide customized patient care based on complex information. Communication with patients has always been taxing in oncology, and applying genomic information in clinical decision making adds to this complexity. Therefore, it is essential to ensure that patients fully understand the role of the testing and how it may affect management decisions. Prior research has reported an insufficient understanding of genomic testing and its applications among those who received such testing for breast cancer susceptibility [15,16,17,18,19]. There is also evidence that patients with cancer have an inadequate understanding of somatic and germline mutations in genomic testing, as well as apprehensiveness towards receiving test results [15,20]. However, studies of patient knowledge and expectations of genomic testing are limited, so it remains unclear how the perspectives of patients and physicians differ regarding the use of this new technology.Most physicians endeavour to provide their patients with enough information about genomic testing and targeted treatments to make informed decisions. However, they do not reliably know how well their patients have assimilated this information. Previous studies of patient and physician expectations of cancer care have revealed significant discrepancies [21]. For example, a survey comparing expectations for cancer survivorship care reported meagre agreement rates [21]. Comparable results were also reported in a similar study published a year later [22]. A multinational survey of oncologists and patients with cancer demonstrated regional variations in the knowledge and clinical applications of personalized medicine [23]. Such discrepancies in patient and physician knowledge and expectations can often result in suboptimal cancer management and follow-up care [20]. Increasing patient–physician communications in areas showing disagreement may lead to better concordance and a higher likelihood of receiving proper care [24,25].With the increasing number and frequency of genomic tests currently being offered to cancer patients, it is essential to ensure that they understand the implications of this new technology. Furthermore, knowledge about the areas of agreement and disagreement between patients and providers can assist in developing strategies for improving education and counselling regarding genomic testing [26,27,28,29,30]. This paper presents the results of a pilot study comparing the self-reported knowledge and expectations regarding genomic testing and identifies differences in their views. 2. Materials and Methods

The study was conducted at the Tom Baker Cancer Center in Alberta, Canada. Two separate but complementary self-administered questionnaires were developed and distributed to 117 patients and 15 medical oncologists over 12 weeks between July and September 2019. The study was designed as a pilot project, with a small number of patients. The Research Ethics Board of the institution approved the study protocol and documents.

The participating oncologists were asked to complete the physician survey after informed consent. Physicians were asked to identify eligible patients based on the study inclusion criteria, which consisted of adult patients who had a primary diagnosis of cancer and had received a diagnostic test to find a targetable mutation. The physicians approached prospective patients during clinic hours, who then completed the survey questionnaire after providing informed consent. The surveys could be completed in the clinic or taken home and returned via email or Canada post.

2.1. Patient QuestionnaireThe patient survey was designed to collect information about their demographics, cancer diagnosis, prior cancer experience, and research interests. It comprised three sections: (a) the patient’s basic knowledge and awareness of cancer and diagnostic tests, (b) the patient’s attitude toward precision oncology and testing, and (c) the patient’s acceptance of integrated testing as a part of disease management, including diagnosis and treatment. The responses to these questions were evaluated in either a binary fashion (“Yes” or “No”) or using a 5-point Likert scale (“Strongly Disagree”, “Disagree”, “Neither Agree nor Disagree”, “Agree”, “Strongly Agree”). The responses were designed to reflect the patient’s understanding, opinion, and acceptance level for genomic testing to help guide their cancer management. For details, please refer to the patient survey in Supplementary File S1. 2.2. Physician QuestionnaireThe physicians were subsequently asked in their survey version to indicate their attitude towards genomic testing, its applications, and how much they used it in their clinical practice. They also identified factors influencing their willingness or disapproval to pursue such testing. Physicians then indicated their perception of the patient’s desire to pursue genomic testing and their concerns about a major or minor procedure undertaken during the testing process. The physician questionnaire consisted of 6 questions, which evaluated their responses on Likert scales similar to those used for the patient survey. For details, please refer to the physician survey in Supplementary File S2. 2.3. Statistical Analysis

Baseline demographics and characteristics for the patient and physician cohorts were summarized with descriptive statistics. Surveys with at least 50% completed responses were included in the data analysis. Each patient’s answers were matched with their corresponding oncologist to form patient–oncologist dyads, and concordance rates were determined. Since the survey responses of the oncologists did not vary from patient to patient, we matched multiple patients to their corresponding oncologists. For patient–oncologist dyads, concordance was defined as a complete agreement in response categories. IBM SPSS 21 statistical software was used for analysis.

In contrast, discordance was signified by any disagreement in response categories between the respondents of a matched survey pair. The patient’s literacy level (at or above high school level) and income (≥CAD 50,000) were also compared with their knowledge of genomic testing using the Pearson chi-squared test. A value of Supplementary File S1). 3. Results(a)

Characteristics of patients and physicians:

Out of 117 surveys distributed among cancer patients, 113 completed surveys were included in the analysis, constituting a 96.4% response rate (Table 1). A total of 15 physicians were approached; 12 were able to participate in the study and were included in the statistical analysis. Among the physicians who participated, 11 physicians were trained medical oncologists, and one was a general practitioner in oncology (GPO). The majority of the physicians were treating at least two tumour sites.In the patient cohort, the study population consisted of 53.9% (n = 61) males and 46.0% (n = 52) females. The median age was 63 yr (range 33–93 yr), and 54.9% (n = 62) of patients had a family history of cancer. We chose different tumour sites to include a diverse patient population. Tumour site distribution among the patients was as follows: 44% (n = 48) breast, 26.5% (n = 30) colon, 4.4% (n = 5) pancreas, 3.5% (n = 4) renal, 9.7% (n = 11) melanoma, 8.8% (n = 10) lung, and 4.4% (n = 5) prostate. Nearly 37% (n = 42) of patients had a university-level education, while 36.3% (n = 41) had college-level degrees or diplomas. While 34.9% (n = 39) of patients had a family member working in healthcare, only 18.6% (n = 21) had healthcare or cancer-related experience. Approximately 70.7% (n = 80) reported an income of >CAD 50,000. Other demographic characteristics are illustrated in Table 1.(b)

Patient understanding of cancer biology and awareness of genomic testing:

Patients were asked questions to examine their understanding of cancer biology, as well as factors contributing to tumour formation, including hereditary, lifestyle, and environments and how genomic testing can help identify tumour characteristics and guide therapy. Overall, the majority (89.4%, n = 101) of patients demonstrated a good understanding of what cancer means and the factors that could contribute to its causation. Out of these, more than 60.0% (n = 70) of patients did not know that genomic changes could lead to tumourgenesis, and only a few (25.7%, n = 29) knew about the role of genomic testing in prognostication. However, over 80.5% (n = 91) of patients understood the uses of genomic testing, including its role in affecting treatment choices and predicting patients’ responses to various treatments (Figure 1a,b). Evaluation of the relationship between the patient literacy and their knowledge of genomic testing showed no significant association, with p values of 0.82 and 0.57, respectively, for Q15 (a) and (b) (please refer to the patient survey and Supplementary Table S1A). Similar results were seen with patients’ income levels and understanding of genomic testing (p = 0.15 and p = 0.09, respectively (please refer to the patient survey and Supplementary Table S1B).(c)

Attitude and expectations of patients towards genomic testing and research:

Patients were also asked questions to understand their comfort level towards receiving unexpected results from genomic testing, which could have health and reproductive implications and impact their decision to pursue the testing [31]. As depicted in Figure 2A,B, 56.6% (n = 64) of patients wanted to undergo testing only for their personal use, and a large number (68.1%, n = 77) were also willing to know about their family’s risk of carrying a mutation that could predispose them to a similar disease. Furthermore, patients were asked about their willingness to biobank their tumour samples for future research, and more than 60.0% (n = 76) consented. While only 28.3% (n = 32) felt that they had enough information to make an informed decision after the informed consent (Figure 2C), more than 60% (n = 72) expressed their willingness to have more formal counselling before pursuing genomic testing for their cancer management (Figure 2D).(d)

Attitudes and expectations of patients vs physicians regarding genomic testing:

We asked patients about their willingness to undergo a major or minor procedure for genomic testing, and this was compared to the physician’s views. Less than 70.0% (n = 78) of patients were willing to undergo minor procedures, which was over-estimated by their physicians in nearly 82.0% (n = 9) of cases. However, many patients felt that their tumours should be tested to guide treatment (70.0%, n = 78), and these views were broadly shared by their oncologists (concordance 65.1%, Table 2).Concerning the factors that influence decision making, the potential to guide treatment influenced both the patients’ and physicians’ decisions the most (78.8%, n = 89 and 63.3%, n = 7, respectively). In addition, physicians were also found to be interested in the ability of genomic testing to predict the disease outcome (63.6%, n = 7). However, the views of patients were different from their oncologists with respect to the goals and expectations of genomic testing. Specifically, patients were more willing to know about their cancer and side effects from treatments and contribute to research, while physicians were more interested in patient outcomes (Figure 3a). Patients also seemed to be the least curious to know about the side effects and disease outcomes when making decisions about testing (Figure 3b).Furthermore, the possibility of a significant complication from the procedure was an essential factor influencing the decision making for both groups (concordance 56.25%). Patients’ and physicians’ views were discrepant regarding genomic testing, causing delays in treatment (concordance 26.7%). This was instead considered an essential factor by the patients to decline genomic testing (41.6, n= 47) in contrast to this consideration for only 18% (n = 2) of physicians (concordance 0.89%). Physicians were more concerned about the impact of genomic testing on the social aspects of the patients’ lives, including the discrimination of health, life, and disability insurance (42.0%, n = 5), in contrast to patients (18.0%, n = 20), who were more worried that the test was of no clinical value (40.0%, n = 45). Importantly, knowledge and expectations regarding the applications of genomic test results on actual diagnosis and prognosis were grossly discrepant between patients and their oncologists (concordance 26.1% and 36.0%, respectively, Table 2). 4. Discussion

The advent of genomic testing capable of identifying clinically significant somatic mutations has revolutionized the world of cancer diagnosis and therapeutics. It helps to characterize tumours at the molecular level, thus enabling medical oncologists to take a more personalized approach to clinical care and optimize patient selection for clinical research. Several new targeted therapies are now widely used and playing critical roles in improving the survival of patients with various cancers. The growing need and application of genomic testing underscore the importance of investigating how well cancer patients understand, value, and assimilate this information.

Patient–physician interaction is an integral part of cancer care delivery, and previous studies have shown that physician communication behavior impacts health outcomes [32,33]. The complexity of precision medicine complicates the exchange of information in significant ways. We found several deficiencies in patients’ understanding of the goal and utilization of genomic testing offered to them. In our study, patients reported dissatisfaction at the end of the informed consent process and demanded more dedicated counselling concerning this topic (Figure 2C,D). In our study, there was no significant correlation found between the patient’s level of education and knowledge of genomic testing, indicating that a patient’s literacy level can sometimes be misleading for the physicians and healthcare teams regarding the amount of information provided to them (Supplementary Table S1A,B). While most of our study patients understood that cancer results from changes in the DNA, known as “mutations,” many patients did not understand the differences between somatic and inherited mutations. Our findings are broadly consistent with previous studies that have also demonstrated patient knowledge gaps concerning their understanding of cancer biology [21,30]. For instance, an analysis of breast cancer patients conducted by Lillie et al. showed that after reading written information about genomic testing, patients with limited health literacy experienced poor recall and were less likely to actively participate in decision making, leading to poor outcomes [34,35]. Therefore, physicians’ lack of awareness about their patients could easily lead to misinterpretation of test results by the patients, as well as create incorrect perceptions about its applications.Therefore, the current informed consent process warrants close examination, since there appears to be a discrepancy between what is conveyed and what is understood by patients. There should be frank, patient-centered, and empathetic discussions aligned with the patient’s wishes and realistic goals [36,37,38]. Such a panel should be modified and simplified to help patients appropriately interpret the test results in their contexts and guide them in making decisions about their care. Emphasis should be placed on using simple language, regardless of their education level, to help patients understand their disease, the role of genomic testing, and its implications. Knowledge checks should supplement this during and at the end of the conversation. Appropriate communication tools, such as patient handouts, videos, or online teaching modules, should be used to deliver complex information to help patients better understand the applications of genomic tests [38]. The use of technology to develop practical communication tools, implement evidence-based decision aids for patients, and create online patient portals to monitor the uptake of such interventions should be promoted [31,38,39].It is also essential to know if the person obtaining the consent has enough knowledge and the required communication skills to share complex information and tailor it to an individual patient’s needs and circumstances [39]. Physicians could also help by ensuring that they feel comfortable having these difficult conversations by educating themselves about the current genomic interventions and their applications. A multilevel interventional approach to patient communication was implemented in the VOICE study [40], showing the tremendous success of multidisciplinary approaches and the value of training physicians in communication skills. We need to invest in interventions focused on patient-centered communication skills training for physicians and other healthcare teams (e.g., genetic counsellors) [32,39]. There is also a role for educating the allied healthcare professionals, including the nurses, with whom patients frequently get in touch to clarify some of the information exchanged during the consultation with their oncologists [39]. Patients with a family history of hereditary cancers should be referred to genetic counsellors, and appropriate guidance should be provided in case of reluctance by involving multidisciplinary teams [38]. In addition, such interventions can be beneficial for physicians, who often lack a background in genetic counselling and testing and are often tasked with interpreting results without assistance [32].Furthermore, one must remember that patients with advanced cancer are often desperate about their prognosis and the diagnostic and treatment choices offered to them [32,38,40]. Consequently, patients often have unrealistic expectations from test results regarding prognosis and cure. In our study, patients were found to overestimate the offered tests’ benefits. On the other hand, the vast majority were willing to learn about cancer risk in their family members (Figure 2B). They believed that genomic information about their cancer could benefit their family members, as well. Still, they did not realize they were also at risk of receiving unexpected results about unknown but clinically relevant oncological mutations. Unfortunately, many of these mutations are rare and do not yet have therapies. Hence, it is crucial to assess each patient’s expectations of personal benefit and benefit to their family members and align them to the treatment goals.