Emergency presentation of colorectal patients in Spain

Methods for improving colorectal cancer annotation efficiency for artificial intelligence-observer training

BACKGROUND

Missing occult cancer lesions accounts for the most diagnostic errors in retrospective radiology reviews as early cancer can be small or subtle, making the lesions difficult to detect. Second-observer is the most effective technique for reducing these events and can be economically implemented with the advent of artificial intelligence (AI).

AIM

To achieve appropriate AI model training, a large annotated dataset is necessary to train the AI models. Our goal in this research is to compare two methods for decreasing the annotation time to establish ground truth: Skip-slice annotation and AI-initiated annotation.

METHODS

We developed a 2D U-Net as an AI second observer for detecting colorectal cancer (CRC) and an ensemble of 5 differently initiated 2D U-Net for ensemble technique. Each model was trained with 51 cases of annotated CRC computed tomography of the abdomen and pelvis, tested with 7 cases, and validated with 20 cases from The Cancer Imaging Archive cases. The sensitivity, false positives per case, and estimated Dice coefficient were obtained for each method of training. We compared the two methods of annotations and the time reduction associated with the technique. The time differences were tested using Friedman's two-way analysis of variance.

RESULTS

Sparse annotation significantly reduces the time for annotation particularly skipping 2 slices at a time (P < 0.001). Reduction of up to 2/3 of the annotation does not reduce AI model sensitivity or false positives per case. Although initializing human annotation with AI reduces the annotation time, the reduction is minimal, even when using an ensemble AI to decrease false positives.

CONCLUSION

Our data support the sparse annotation technique as an efficient technique for reducing the time needed to establish the ground truth.

The Oddity Detection in Diverse Scenes (ODDS) database: Validated real-world scenes for studying anomaly detection

Many applied screening tasks (e.g., medical image or baggage screening) involve challenging searches for which standard laboratory search is rarely equivalent. For example, whereas laboratory search frequently requires observers to look for precisely defined targets among isolated, non-overlapping images randomly arrayed on clean backgrounds, medical images present unspecified targets in noisy, yet spatially regular scenes. Those unspecified targets are typically oddities, elements that do not belong. To develop a closer laboratory analogue to this, we created a database of scenes containing subtle, ill-specified "oddity" targets. These scenes have similar perceptual densities and spatial regularities to those found in expert search tasks, and each includes 16 variants of the unedited scene wherein an oddity (a subtle deformation of the scene) is hidden. In Experiment 1, eight volunteers searched thousands of scene variants for an oddity. Regardless of their search accuracy, they were then shown the highlighted anomaly and rated its subtlety. Subtlety ratings reliably predicted search performance (accuracy and response times) and did so better than image statistics. In Experiment 2, we conducted a conceptual replication in which a larger group of naïve searchers scanned subsets of the scene variants. Prior subtlety ratings reliably predicted search outcomes. Whereas medical image targets are difficult for naïve searchers to detect, our database contains thousands of interior and exterior scenes that vary in difficulty, but are nevertheless searchable by novices. In this way, the stimuli will be useful for studying visual search as it typically occurs in expert domains: Ill-specified search for anomalies in noisy displays.

Presentation Patterns and Outcomes in Patients with Colorectal Cancer Seeking the Emergency Department for Consultation

Background Patients with colorectal cancer may seek the emergency department for symptoms related to chemotherapy and radiotherapy side effects as well as those from the disease itself and from surgery complications.

Objectives To establish the epidemiological and clinical profile of colorectal cancer patients that look for consultations in the emergency department.

Methods Retrospective study of emergency room charts from colorectal cancer patients that consulted in a single oncological hospital for the period of 1 year.

Results Four hundred and forty-six consultations were identified (49.5% males and 50.5% females) with a mean age of 63 years and with advanced disease (most with tumor, node, metastases [TNM] stages III and IV). The most common complaint was abdominal pain (27.5%), followed by nausea (4.7%; more commonly seen in females with p = 0.03) and bladder symptoms (4.7%; more commonly seen in males, with p = 0.003). Infections (10.3%) and acute abdominal pain (9.1%) were the most frequent diagnoses. About 18% of them were admitted to the hospital and 80% were discharged home.

Conclusion The profile of patients with colorectal cancer seeking the emergency department comprises patients with advanced disease and a similar proportion of males and females. Symptom-driven complaints were the most frequent reason for consultations.

Relationship between time from symptom's onset to diagnosis and prognosis in patients with symptomatic colorectal cancer

BACKGROUND

Controversy exists regarding the relationship of the outcome of patients with colorectal cancer (CRC) with the time from symptom onset to diagnosis. The aim of this study is to investigate this association, with the assumption that this relationship was nonlinear and with adjustment for multiple confounders, such as tumor grade, symptoms, or admission to an emergency department.

METHODS

This multicenter study with prospective follow-up was performed in five regions of Spain from 2010 to 2012. Symptomatic cases of incident CRC from a previous study were examined. At the time of diagnosis, each patient was interviewed, and the associated hospital and clinical records were reviewed. During follow-up, the clinical records were reviewed again to assess survival. Cox survival analysis with a restricted cubic spline was used to model overall and CRC-specific survival, with adjustment for variables related to the patient, health service, and tumor.

RESULTS

A total of 795 patients had symptomatic CRC and 769 of them had complete data on diagnostic delay and survival. Univariate analysis indicated a lower HR for death in patients who had diagnostic intervals less than 4.2 months. However, after adjustment for variables related to the patient, tumor, and utilized health service, there was no relationship of the diagnostic delay with survival of patients with colon and rectal cancer, colon cancer alone, or rectal cancer alone. Cubic spline analysis indicated an inverse association of the diagnostic delay with 5-year survival. However, this association was not statistically significant.

CONCLUSIONS

Our results indicated that the duration of diagnostic delay had no significant effect on the outcome of patients with CRC. We suggest that the most important determinant of the duration of diagnostic delay is the biological profile of the tumor. However, it remains the responsibility of community health centers and authorities to minimize diagnostic delays in patients with CRC and to implement initiatives that improve early diagnosis and provide better outcomes.

Real-world colorectal cancer diagnostic pathways in Ontario, Canada: A population-based study

OBJECTIVE

This study aimed to identify colorectal cancer (CRC) diagnostic pathways and describe patients in those pathway groups.

METHODS

This was a cross-sectional study of CRC patients in Ontario, Canada, diagnosed 2009-2012 that used linked administrative data at ICES. We used cluster analysis on 11 pathway variables characterising patient presentation, symptoms, procedures and referrals. We assessed associations between patient- and disease-related characteristics and diagnostic pathway group. We further characterised the pathways by diagnostic interval and number of related physician visits.

RESULTS

Six diagnostic pathways were identified, with three adhering to provincial diagnostic guidelines: screening (N = 4494), colonoscopy (N = 10,066) and imaging plus colonoscopy (N = 3427). Non-adherent pathways were imaging alone (N = 2238), imaging and emergency presentation (N = 2849) and no pre-diagnostic workup (N = 887). Patients in adherent pathways were younger, had fewer comorbidities, lived in less deprived areas and had earlier stage disease. The median diagnostic interval length varied across pathways from 12 to 126 days, correlating with the number of CRC-related visits.

CONCLUSIONS

This study demonstrated substantial variations in real-world CRC diagnostic pathways and 25% were diagnosed through non-adherent pathways. Those patients were older, had more comorbid disease and had higher stage cancer. Further research needs to identify and describe the reasons for divergent diagnostic processes.

Patterns of emergency department visits preceding colorectal cancer diagnosis: a population-based study

Aim: To assess the patterns of emergency department (ED) visits in the 3 months preceding a diagnosis of colorectal cancer (CRC) in a real-world, population-based context. Materials & methods: Linked provincial registries in Alberta, Canada, were accessed and patients with CRC diagnosed between 2004 and 2018 were identified. The National Ambulatory Care reporting system was used to identify patients who visited an ED within 3 months of a diagnosis of CRC. Multivariable logistic regression analysis was used to identify factors associated with any ED visits as well as frequent (≥3) ED visits. Results: A total of 25,310 patients with CRC were included in the current study. These include 10,126 patients (40%) who had at least one visit to the ED in the 3 months before a diagnosis of CRC diagnosis and 613 patients (2.4%) who visited the ED multiple (≥3) times. The following factors were associated with any visit to an ED: older age (odds ratio [OR]: 1.010; 95% CI: 1.008-1.012), female gender (OR: 1.23; 95% CI: 1.16-1.30), higher comorbidity index (OR: 1.38; 95% CI: 1.35-1.41), metastatic disease (OR: 2.37; 95% CI: 2.23-2.53), proximal tumors (OR: 1.59; 95% CI: 1.50-1.68) and North zone (OR vs south zone: 1.75; 95% CI: 1.55-1.98). Conclusion: It is not uncommon for CRC patients to visit the ED at least once in the 3 months prior to having such a diagnosis. Factors associated with frequent pre diagnosis emergency visits included female gender, higher burden of comorbid disease, advanced stage, proximal tumors and living in the North zone of Alberta (where there is limited access to specialist care).

Impact of a colorectal cancer screening program implantation on delays and prognosis of non-screening detected colorectal cancer

BACKGROUND

The implementation of a colorectal cancer (CRC) screening programme may increase the awareness of Primary Care Physicians, reduce the diagnostic delay in CRC detected outside the scope of the screening programme and thus improve prognosis.

AIM

To determine the effect of implementation of a CRC screening programme on diagnostic delays and prognosis of CRC detected outside the scope of a screening programme.

METHODS

We performed a retrospective intervention study with a pre-post design. We identified 322 patients with incident and confirmed CRC in the pre-implantation cohort (June 2014 – May 2015) and 285 in the post-implantation cohort (June 2017 - May 2018) in the Cancer Registry detected outside the scope of a CRC screening programme. In each patient we calculated the different healthcare diagnostics delays: global, primary and secondary healthcare, referral and colonoscopy-related delays. In addition, we collected the initial healthcare that evaluated the patient, the home location (urban/rural), and the CRC stage at diagnosis. We determined the two-year survival and we performed a multivariate proportional hazard regression analysis to determine the variables associated with survival.

RESULTS

We did not detect any differences in the patient or CRC baseline-related variables. A total of 20.1% of patients was detected with metastatic disease. There was a significant increase in direct referral to colonoscopy from primary healthcare (25.5%, 35.8%; P = 0.04) in the post-implantation cohort. Diagnostic delay was reduced by 24 d (106.64 ± 148.84 days, 82.84 ± 109.31 d; P = 0.02) due to the reduction in secondary healthcare delay (46.01 ± 111.65 d; 29.20 ± 60.83 d; P = 0.02). However, we did not find any differences in CRC stage at diagnosis or in two-year survival (70.3%; P = 0.9). Variables independently associated with two-year risk of death were age (Hazard Ratio-HR: 1.06, 95%CI: 1.04-1.07), CRC stage (II HR: 2.17, 95%CI: 1.07-4.40; III HR: 3.07, 95%CI: 1.56-6.08; IV HR: 19.22, 95%CI: 9.86-37.44; unknown HR: 9.24, 95%CI: 4.27-19.99), initial healthcare consultation (secondary HR: 2.93, 95%CI: 1.01-8.55; emergency department HR: 2.06, 95%CI: 0.67-6.34), hospitalization during the diagnostic process (HR: 1.67, 95%CI: 1.17-2.38) and urban residence (HR: 1.44, 95%CI: 1.06-1.98).

CONCLUSION

Although implementation of a CRC screening programme can reduce diagnostic delays for CRC detected in symptomatic patients, this has no effect on CRC stage or survival.

How best to palliate and treat emergency conditions in geriatric patients with colorectal cancer

Almost one third of colorectal cancer (CRC) cases are diagnosed in an emergency setting, mostly among geriatric patients. Clinical scenarios are often complex and decision making delicate. Besides the obvious need to consider the patient's and/or family and care givers' desires, the surgeon should be able to make the best educated guess on future outcomes in three areas: oncological prognosis, morbidity and mortality risk, and long-term functional loss. Using simple and brief tools for frailty screening reasonable treatment goals with curative or palliative intent can be planned. The most frequent clinical scenarios of CRC in emergency are bowel obstruction and perforation. We propose treatment algorithms based on assessment of the patient's overall reserve and discuss the indications, techniques and impact of a stoma in the geriatric patient. Bridge to surgery strategies may be best adapted to help the frail geriatric patient overcome the acute disease and maybe return to previous state of function. Post-operative morbidity and mortality rates are high in emergency surgery for CRC, but if the geriatric patient survives the post-operative period, oncological prognosis seems to be similar to younger patients. Because the occurrence of complications is the strongest predictor of functional decline and death, post-operative care plays a major role to optimize outcomes. Future studies should further investigate emergency surgery of CRC in the older adults focusing in particular on functional outcomes in order to help physicians counsel patients and families for a tailored treatment.

Correction: Emergency presentation of colorectal patients in Spain

[This corrects the article DOI: 10.1371/journal.pone.0203556.].