Distance Traveled by Patients Globally to Access Radiotherapy: A Systematic Review

PURPOSE/OBJECTIVE(S)

This systematic review aims to investigate the global travel patterns of patients seeking radiotherapy and examines the distance traveled by patients and its impact on secondary outcomes such as travel time and survival. The findings of this review will provide crucial information on barriers to accessing radiotherapy and inform the development of patient-centered care strategies aimed at improving access to this important form of treatment.

MATERIALS/METHODS

A comprehensive search of four databases was conducted from June to August 2022. Studies were included if they were observational, retrospective, or randomized/non-randomized, published between June 2000 and June 2022, and reported the distance traveled globally for the treatment of malignant or benign disease. Studies were excluded if they did not report travel distance or were not written in English.

RESULTS

A total of 176 studies were included. Most of the studies (69.9%) were conducted in North America, with the majority (68.7%) in the United States. The treatment modalities varied with external beam radiation therapy being the most common (17.0%). The most common disease site was breast (26.7%). Of the included studies, 49 reported the mean distance traveled for radiation therapy. The shortest mean distance was reported in the United States at 4.83 miles, while the longest was reported in Iran at 276.5 miles. It was observed that patients living in countries outside the United States traveled greater distances for radiation therapy than those living within the U.S. Additional factors such as urban vs. rural residence and treatment modality were also found to impact the distance traveled for radiation therapy. Our results indicate a wide range of travel times, with approximately half of the studies reporting values greater than 1 hour, which was the case for 100% of the studies on low-income populations (n = 4). Out of 176 studies, only 15% discussed patient survival and reported conflicting results between travel distance and survival rates, regardless of treatment, disease site, or country of origin.

CONCLUSION

This systematic review is the most comprehensive to date on the global travel patterns of patients seeking radiotherapy. Results show that travel distances varied, but overall, patients in the U.S. traveled shorter distances for radiation therapy than those living outside the country. Treatment center location, patient residence, and treatment modality impacted patient travel distance, but the patterns were inconsistent. These findings emphasize the importance of considering the distance traveled as a barrier to receiving radiotherapy and highlight the need for strategies to improve patient access and prioritize patient-centered care.

Machine Learning for the Prediction of Biochemical Recurrence in Patients Treated with Radical Prostatectomy

PURPOSE/OBJECTIVE(S)

Biochemical recurrence (BCR) occurs in about 40% of patients with prostate cancer following radical prostatectomy (RP). Our goal was to develop a machine learning model for the prediction of BCR five-years after RP, to improve patient prognostication.

MATERIALS/METHODS

Patients treated with RP at a tertiary care medical center between 1990 and 2017 were included. A gradient boosted decision trees-based machine learning model modified to handle survival data was trained on 80% of the dataset. The model's performance was evaluated on the remaining 20%. Input variables were age at surgery, prostate specific antigen (PSA) at diagnosis (in ng/mL), pathologic Gleason grade group (GG), pathologic T stage (organ confined disease vs. extracapsular extension (ECE) vs. seminal vesicle invasion (SVI)), lymph node involvement, and surgical margin status. Model performance was assessed using time-dependent area under curve of the receiver operator curve (AUC).

RESULTS

The full dataset included 11,139 patients, of whom 1,153 (10%) developed BCR. Median age at surgery was 59 and PSA at diagnosis was 5.4 ng/mL. Only 1,080 (9.7%) patients had GG 3, and 707 (6.3%) GG 4 and 5. 1,366 (12%) patients had positive surgical margins and 134 (1.2%) had lymph node involvement. Most patients had organ confined disease with EPE and SVI diagnosed in 2,759 (25%) and 392 (3.5%) patients, respectively. Median follow-up was 5 years and median time to BCR was 4 years. When validated on the hold-out set of 2,228 patients, the model shows a time-dependent AUC of 0.82 (95% CI 0.78 - 0.86) for BCR at t = 5 years.

CONCLUSION

Our machine learning model can be used to estimate risk of BCR following RP and shows exceptional performance, with implications on patient prognostication and follow-up. We are currently working on validating its performance on an external dataset.

Identifying Common Topics in Patient Portal Messages with Unsupervised Natural Language Processing

PURPOSE/OBJECTIVE(S)

Patient portal messaging is an increasingly important form of communication between patients and medical providers. This has become particularly relevant in oncology, where patients undergo intense longitudinal treatments that require frequent communication regarding symptoms, appointments, and diagnostic results. The rise in the volume of these messages has significantly increased the workload of medical providers and consequent physician burn-out. Natural language processing (NLP), particularly transformer-based models, may offer an automated approach to characterize the content of patient messages and improve message triage and routing. In this study, we employed a state-of-the-art language model (Bidirectional Encoder Representations from Transformers; BERT) to identify data-derived categories of representative topics from real-world data thereby providing basic information to build an appropriate routing system.

MATERIALS/METHODS

Patient-generated portal messages sent to a messaging pool for a single institution radiation oncology department from 2014 to 2023 were extracted. BERTopic, an NLP-based topic modeling technique based on BERT was optimized for topic modeling of patient messages. Uniform Manifold Approximation and Projection (UMAP) was used to reduce dimensionality and visualize topic relationships across messages. The BERTopic-identified topic categories were subsequently labeled manually by one of the physician investigators. Differences of number of messages over time were assessed using t-tests.

RESULTS

A total of 47,492 messages were retrieved. The average number of messages per month from a single patient ranged from 1 to 18 (median 1.67, interquartile range 1.0-2.4). The total volume of patient messages showed a ten-fold increase over the study period, with 101 messages per month sent in 2014 and 999 messages per month in 2022 (p<0.001). BERTopic initially identified 35 topics whose relationships and degrees of overlap were visualized by UMAP. Due to physician-identified similarities, these topics were reduced into 13 categories. The most frequent topic category was messages about laboratory tests or imaging studies: 24.3%, followed by messages expressing appreciation: 18.9%, scheduling discussions: 15.6%, symptom-related messages: 11%, and treatment-related messages: 10.7%.

CONCLUSION

Patient portal messages sent to a single institution radiation oncology department have increased dramatically in volume since implementation, corresponding to a broader national trend. NLP successfully identified common subject themes across patient messages, many of which are related to scheduling. This presents potential opportunities to apply NLP to automate message routing in the future.

Impacts of Wildfire Events on California Radiation Oncology Clinics and Patients

PURPOSE/OBJECTIVE(S)

The impact of climate-related disasters such as wildfires on healthcare delivery and cancer care is a growing concern. Patients undergoing radiotherapy are particularly vulnerable to treatment interruptions which are known to have a direct impact on survival outcomes. We report results of the first pilot study characterizing the impact of wildfires on radiation oncology clinics and their patients.

MATERIALS/METHODS

A survey was sent to 415 California radiation oncologists representing 144 clinics identified using the ASTRO member directory to gather information about clinic and radiation oncologist demographics, wildfires' impacts on the clinic (physical/operational), physicians, staff, and patients between 2017 and 2022, as well as clinics' disaster preparedness efforts.

RESULTS

A total of 51 radiation oncologists completed the survey, representing 43 clinics (30% of clinics) in 24 (41%) of California counties. 35 (69%) of respondents self-identified as male, 27 (53%) worked at hospital-affiliated centers, 19 (37%) worked in academic centers, with 49 (96%) practicing in metropolitan areas. A total of 31 clinics were impacted by a wildfire between 2017 and 2020. The two rural clinics (100%) and 29 (59%) of metro practices reported being impacted by wildfires in the last 5 years, with 9 (18%) reporting a clinic closure and 15 (29%) reporting staffing shortages. 28 (55%) of respondents reported impacts on patients including having to evacuate, 27 (53%) having to cancel or reschedule treatments, and 23 (45%) experiencing physical, mental, or financial hardship due to the wildfires. Among the 25 clinics impacted by wildfires, 12 (24%) reported physical/operational impacts including being forced to evacuate patients to another treatment center, transportation interruptions (19, 37%), community and regional evacuations (18, 35%), school closures (18, 35%), and physical/mental health impacts (14, 27%) on staff due to the wildfires. Small clinics (25 staff or less) that experienced a wildfire were twice as likely to experience closures (6 of 16 clinics, 38%) but equally likely to experience staffing shortages (8 of 16, 50 %) as compared to larger practices (7 of 15, 47%). Nearly half of respondents 25 (47%) reported their workplace had a wildfire emergency preparedness plan.

CONCLUSION

The results of this study demonstrate the significant impact wildfires have on patient care in both rural and metropolitan areas. The findings emphasize the importance of emergency preparedness planning to minimize the consequences of such disasters and underscores the need for further research to explore risk factors associated with patient and community vulnerability to climate-related crises. Such research will be essential to informing and developing future emergency preparedness plans.

The Environmental Impact of Radiation Oncology: The "Footprint" of External Beam Radiation Therapy

PURPOSE/OBJECTIVE(S)

There is a growing concern for the healthcare sector's impact on the environment. Prior carbon impact studies in radiation oncology have been limited in scope and methodology. This study aims to fill this gap by using an internationally recognized cradle-to-grave life cycle assessment (LCA) approach to quantify all environmental impacts from raw material extraction to product disposal for external beam radiation therapy (EBRT) in treating the most commonly diagnosed cancers.

MATERIALS/METHODS

This LCA was performed in accordance with ISO 14040 and 14044 at a single academic medical center. It quantified the environmental impact of EBRT across four categories: global warming potential (GWP), carcinogenic and non-carcinogenic human toxicity, and respiratory effects (PM2.5), from initial consultation to the completion of the last EBRT fraction for each disease site. Data collection involved weighing all materials used, measuring/calculating building and equipment electricity usage (e.g., HVAC and Linacs), and recording patient and staff transit. The study analyzed the impact of both minimum and maximum fractionations for each disease site and simulated alternative clinical scenarios such as telemedicine, renewable energy use and hypofractionation.

RESULTS

Regardless of disease site, there were significant differences in the environmental impacts associated with transit, electricity and supplies for EBRT treatment cycles. Staff and patient transport contributed the most, accounting for >92% of the total environmental impact including GWP (5.02x102 ± 9.38x101 kgCO2eq), carcinogenic (6.25x10-5 ± 1.23x10-5 CTUh) and non-carcinogenic human toxicity (1.16x10-4 ± 2.35x10-5 CTUh). Electricity accounted for 1-13% of the total impact, with most impact arising from respiratory effects (3.05x10-2 kg ± 2.72x10-3 PM2.5). The impact of supplies and materials was less than 3% across all categories. Alternative scenario modeling showed that telemedicine had a maximum impact reduction of 3.5% (2.54x 101kgCO2eq) for GWP, while renewable energy use had a maximum impact reduction of 8% (2.37 x 10-2 PM2.5) for respiratory effects. Reducing the number of total treatment days via hypofractionation can reduce GWP by 67-78% and carcinogenic emissions by 63-77% (3.48 x 102 - 5.53 x 102 kgCO2eq) and (3.73 x 10-5 - 6.85 x 10-5CTUh), respectively, with variation depending on the total number of fractions.

CONCLUSION

This study provides a comprehensive environmental impact assessment for EBRT among the most commonly treated disease sites, establishing a baseline metric and identifying targets for impact reduction. We are currently performing a multi-center validation study to be completed by June 2023. Our findings fill an important gap in cancer care and are critical for developing sustainable practices in the face of increasing demand for radiotherapy in a changing climate. LCAs evaluating all aspects of cancer care will be essential for promoting equitable and sustainable care.

Machine Learning for the Prediction of Adverse Pathological Outcomes in Patients Treated with Radical Prostatectomy

PURPOSE/OBJECTIVE(S)

Extracapsular extension (ECE) and seminal vesicle invasion (SVI) are associated with negative oncologic outcomes in patients with prostate cancer. We have developed and validated a machine learning model to more accurately identify patients at risk of these adverse surgical outcomes prior to radical prostatectomy (RP).

MATERIALS/METHODS

This study included a cohort of patients diagnosed with prostate cancer and treated with RP and lymph node dissection at a tertiary care medical center from 2010 to 2020. An ensemble model using a base gradient-boosted trees-based machine learning model and isotonic calibrators was trained on 80% of the cohort, with 20% held out for validation. The model uses age at surgery, prostate specific antigen level (PSA) at diagnosis, biopsy Gleason grade group, numbers of positive and negative cores on biopsy, and clinical T-stage (cT) as input variables. Model performance was assessed on the hold-out set using the area under the receiver operating curve (AUC).

RESULTS

The full dataset included 18,729 eligible patients. Median PSA at diagnosis was 7.3 ng/mL. Most patients had clinically organ confined disease (cT1 - cT2) with only 136 (0.7%) having cT3. The most common biopsy Gleason grade group was 2 (7,118 or 38% of patients), with Gleason grade 4 in 1,796 (9.6%), and 5 in 1,064 (5.7%) patients. After RP, 11,931 (64%) of patients had organ confined disease, 4,298 (23%) had ECE, and 2,500 (13%) had SVI. When validated on the hold-out set (n = 3,746), the model had AUCs of 0.79 (95%-CI 0.77 - 0.80), 0.67 (0.65 - 0.69), and 0.83 (0.81 - 0.85) for the prediction of organ confined disease, ECE, and SVI, respectively.

CONCLUSION

In conclusion, we have developed a machine learning model that predicts individual patient risk of pathologic T-stage. The model can be used to provide more accurate risk assessments and improve surgical treatment planning. We are currently working on externally validating our results on patients from different institutions.

Inheritance of kernel elongation in F2 rice populations crossed between Indian rice basmati 370 and selected Malaysian rice varieties

The speciality rice variety such as Basmati and Jasmine maybe not be very suitable to plant under the Malaysian climate because the growth may be influenced by the changes in the environmental conditions causing the changes of its physical and chemical characters and results in low yield. Thus, an attempt was made through a hybridisation program among selected Malaysian rice varieties consist of MRQ50, MRQ74, MRQ76, MR219, Mahsuri Mutant, and Mahsuri Mutant 98 with Indian traditional rice, Basmati 370 with the main objectives, to compare the kernel elongation ratio of each cross and to determine the potential and suitable combination among crosses rice for further evaluation program. The results of the study indicated that among parental lines, Basmati 370 posed the highest elongation ratio 2.00 which revealed its speciality elongation trait of cooked rice. A comparison between crosses, rice cross between Basmati 370 and Mahsuri Mutant 98 showed the highest elongation ratio with the value 2.12 and the lowest elongation ratio of about 1.96 was found in rice cross between Basmati 370 and MR219. However, all hybridised rice poses a value of elongation ratio more than 1.6 which is considered good kernel elongation. The segregation pattern of all crosses also posed a ratio 3:1 whereas three is low elongation and one is high elongation which demonstrated the Mendelian inheritance of monogenic cross. The results obtained from this study could contribute to the existing literature on speciality rice production in Malaysia and beneficial for future rice quality improvement program.

Acute and delayed protective effects of pharmacologically induced hypothermia in an intracerebral hemorrhage stroke model of mice

Hemorrhagic stroke, including intracerebral hemorrhage (ICH), is a devastating subtype of stroke; yet, effective clinical treatment is very limited. Accumulating evidence has shown that mild to moderate hypothermia is a promising intervention for ischemic stroke and ICH. Current physical cooling methods, however, are less efficient and often impractical for acute ICH patients. The present investigation tested pharmacologically induced hypothermia (PIH) using the second-generation neurotensin receptor (NTR) agonist HPI-201 (formerly known as ABS-201) in an adult mouse model with ICH. Acute or delayed administrations of HPI-201 (2mg/kg bolus injection followed by 2 injections of 1mg/kg, i.p.) were initiated at 1 or 24h after ICH. HPI-201 induced mild hypothermia within 30 min and body and brain temperatures were maintained at 32.7 ± 0.4°C for at least 6h without causing observable shivering. With the 1-h delayed treatment, HPI-201-induced PIH significantly reduced ICH-induced cell death and brain edema compared to saline-treated ICH animals. When HPI-201-induced hypothermia was initiated 24h after the onset of ICH, it still significantly attenuated brain edema, cell death and blood-brain barrier breakdown. HPI-201 significantly decreased the expression of matrix metallopeptidase-9 (MMP-9), reduced caspase-3 activation, and increased Bcl-2 expression in the ICH brain. Moreover, ICH mice received 1-h delayed HPI-201 treatment performed significantly better in the neurological behavior test 48 h after ICH. All together, these data suggest that systemic injection of HPI-201 is an effective hypothermic strategy that protects the brain from ICH injury with a wide therapeutic window. The protective effect of this PIH therapy is partially mediated through the alleviation of apoptosis and neurovascular damage. We suggest that pharmacological hypothermia using the newly developed neurotensin analogs is a promising therapeutic treatment for ICH.