Article Type
Changed
Tue, 08/20/2024 - 15:56

 

TOPLINE:

A novel method combining genetic variants, symptoms, and patient characteristics is moderately successful at predicting which primary care patients are at high risk of developing colorectal cancer (CRC) within the next 2 years — without the benefit of a faecal immunochemical test (FIT). Up to 16% of primary care patients are non-compliant with FIT, which is the gold standard for predicting CRC.

METHODOLOGY:

  • This study was retrospective cohort of 50,387 UK Biobank participants reporting a CRC symptom in primary care at age ≥ 40 years.
  • The novel method, called an integrated risk model, used a combination of a polygenic risk score from genetic testing, symptoms, and patient characteristics to identify patients likely to develop CRC in the next 2 years.
  • The primary outcome was the risk model’s performance in classifying a CRC case according to a statistical metric, the receiver operating characteristic area under the curve. Values range from 0 to 1, where 1 indicates perfect discriminative power and 0.5 indicates no discriminative power.

TAKEAWAY:

  • The cohort of 50,387 participants was found to have 438 cases of CRC and 49,949 controls without CRC within 2 years of symptom reporting. CRC cases were diagnosed by hospital records, cancer registries, or death records.
  • Increased risk of a CRC diagnosis was identified by a combination of six variables: older age at index date of symptom, higher polygenic risk score, which included 201 variants, male sex, previous smoking, rectal bleeding, and change in bowel habit.
  • The polygenic risk score alone had good ability to distinguish cases from controls because 1.45% of participants in the highest quintile and 0.42% in the lowest quintile were later diagnosed with CRC.
  • The variables were used to calculate an integrated risk model, which estimated the cross-sectional risk (in 80% of the final cohort) of a subsequent CRC diagnosis within 2 years. The highest scoring integrated risk model in this study was found to have a receiver operating characteristic area under the curve value of 0.76 with a 95% CI of 0.71-0.81. (A value of this magnitude indicates moderate discriminative ability to distinguish cases from controls because it falls between 0.7 and 0.8. A higher value [above 0.8] is considered strong and a lower value [< 0.7] is considered weak.)

IN PRACTICE:

The authors concluded, “The [integrated risk model] developed in this study predicts, with good accuracy, which patients presenting with CRC symptoms in a primary care setting are likely to be diagnosed with CRC within the next 2 years.”

The integrated risk model has “potential to be immediately actionable in the clinical setting … [by] inform[ing] patient triage, improving early diagnostic rates and health outcomes and reducing pressure on diagnostic secondary care services.”

SOURCE:

The corresponding author is Harry D. Green of the University of Exeter, England. The study (2024 Aug 1. doi: 10.1038/s41431-024-01654-3) appeared in the European Journal of Human Genetics.

LIMITATIONS:

Limitations included an observational design and the inability of the integrated risk model to outperform FIT, which has a receiver operating characteristic area under the curve of 0.95.

DISCLOSURES:

None of the authors reported competing interests. The funding sources included the National Institute for Health and Care Research and others.

A version of this article first appeared on Medscape.com.

Publications
Topics
Sections

 

TOPLINE:

A novel method combining genetic variants, symptoms, and patient characteristics is moderately successful at predicting which primary care patients are at high risk of developing colorectal cancer (CRC) within the next 2 years — without the benefit of a faecal immunochemical test (FIT). Up to 16% of primary care patients are non-compliant with FIT, which is the gold standard for predicting CRC.

METHODOLOGY:

  • This study was retrospective cohort of 50,387 UK Biobank participants reporting a CRC symptom in primary care at age ≥ 40 years.
  • The novel method, called an integrated risk model, used a combination of a polygenic risk score from genetic testing, symptoms, and patient characteristics to identify patients likely to develop CRC in the next 2 years.
  • The primary outcome was the risk model’s performance in classifying a CRC case according to a statistical metric, the receiver operating characteristic area under the curve. Values range from 0 to 1, where 1 indicates perfect discriminative power and 0.5 indicates no discriminative power.

TAKEAWAY:

  • The cohort of 50,387 participants was found to have 438 cases of CRC and 49,949 controls without CRC within 2 years of symptom reporting. CRC cases were diagnosed by hospital records, cancer registries, or death records.
  • Increased risk of a CRC diagnosis was identified by a combination of six variables: older age at index date of symptom, higher polygenic risk score, which included 201 variants, male sex, previous smoking, rectal bleeding, and change in bowel habit.
  • The polygenic risk score alone had good ability to distinguish cases from controls because 1.45% of participants in the highest quintile and 0.42% in the lowest quintile were later diagnosed with CRC.
  • The variables were used to calculate an integrated risk model, which estimated the cross-sectional risk (in 80% of the final cohort) of a subsequent CRC diagnosis within 2 years. The highest scoring integrated risk model in this study was found to have a receiver operating characteristic area under the curve value of 0.76 with a 95% CI of 0.71-0.81. (A value of this magnitude indicates moderate discriminative ability to distinguish cases from controls because it falls between 0.7 and 0.8. A higher value [above 0.8] is considered strong and a lower value [< 0.7] is considered weak.)

IN PRACTICE:

The authors concluded, “The [integrated risk model] developed in this study predicts, with good accuracy, which patients presenting with CRC symptoms in a primary care setting are likely to be diagnosed with CRC within the next 2 years.”

The integrated risk model has “potential to be immediately actionable in the clinical setting … [by] inform[ing] patient triage, improving early diagnostic rates and health outcomes and reducing pressure on diagnostic secondary care services.”

SOURCE:

The corresponding author is Harry D. Green of the University of Exeter, England. The study (2024 Aug 1. doi: 10.1038/s41431-024-01654-3) appeared in the European Journal of Human Genetics.

LIMITATIONS:

Limitations included an observational design and the inability of the integrated risk model to outperform FIT, which has a receiver operating characteristic area under the curve of 0.95.

DISCLOSURES:

None of the authors reported competing interests. The funding sources included the National Institute for Health and Care Research and others.

A version of this article first appeared on Medscape.com.

 

TOPLINE:

A novel method combining genetic variants, symptoms, and patient characteristics is moderately successful at predicting which primary care patients are at high risk of developing colorectal cancer (CRC) within the next 2 years — without the benefit of a faecal immunochemical test (FIT). Up to 16% of primary care patients are non-compliant with FIT, which is the gold standard for predicting CRC.

METHODOLOGY:

  • This study was retrospective cohort of 50,387 UK Biobank participants reporting a CRC symptom in primary care at age ≥ 40 years.
  • The novel method, called an integrated risk model, used a combination of a polygenic risk score from genetic testing, symptoms, and patient characteristics to identify patients likely to develop CRC in the next 2 years.
  • The primary outcome was the risk model’s performance in classifying a CRC case according to a statistical metric, the receiver operating characteristic area under the curve. Values range from 0 to 1, where 1 indicates perfect discriminative power and 0.5 indicates no discriminative power.

TAKEAWAY:

  • The cohort of 50,387 participants was found to have 438 cases of CRC and 49,949 controls without CRC within 2 years of symptom reporting. CRC cases were diagnosed by hospital records, cancer registries, or death records.
  • Increased risk of a CRC diagnosis was identified by a combination of six variables: older age at index date of symptom, higher polygenic risk score, which included 201 variants, male sex, previous smoking, rectal bleeding, and change in bowel habit.
  • The polygenic risk score alone had good ability to distinguish cases from controls because 1.45% of participants in the highest quintile and 0.42% in the lowest quintile were later diagnosed with CRC.
  • The variables were used to calculate an integrated risk model, which estimated the cross-sectional risk (in 80% of the final cohort) of a subsequent CRC diagnosis within 2 years. The highest scoring integrated risk model in this study was found to have a receiver operating characteristic area under the curve value of 0.76 with a 95% CI of 0.71-0.81. (A value of this magnitude indicates moderate discriminative ability to distinguish cases from controls because it falls between 0.7 and 0.8. A higher value [above 0.8] is considered strong and a lower value [< 0.7] is considered weak.)

IN PRACTICE:

The authors concluded, “The [integrated risk model] developed in this study predicts, with good accuracy, which patients presenting with CRC symptoms in a primary care setting are likely to be diagnosed with CRC within the next 2 years.”

The integrated risk model has “potential to be immediately actionable in the clinical setting … [by] inform[ing] patient triage, improving early diagnostic rates and health outcomes and reducing pressure on diagnostic secondary care services.”

SOURCE:

The corresponding author is Harry D. Green of the University of Exeter, England. The study (2024 Aug 1. doi: 10.1038/s41431-024-01654-3) appeared in the European Journal of Human Genetics.

LIMITATIONS:

Limitations included an observational design and the inability of the integrated risk model to outperform FIT, which has a receiver operating characteristic area under the curve of 0.95.

DISCLOSURES:

None of the authors reported competing interests. The funding sources included the National Institute for Health and Care Research and others.

A version of this article first appeared on Medscape.com.

Publications
Publications
Topics
Article Type
Sections
Disallow All Ads
Content Gating
No Gating (article Unlocked/Free)
Alternative CME
Disqus Comments
Default
Use ProPublica
Hide sidebar & use full width
render the right sidebar.
Conference Recap Checkbox
Not Conference Recap
Clinical Edge
Display the Slideshow in this Article
Medscape Article
Display survey writer
Reuters content
Disable Inline Native ads
WebMD Article