A 55-year-old man is evaluated in the urology clinic for an elevated prostate-specific ant… — USMLE Step 3

A 55-year-old man is evaluated in the urology clinic for an elevated prostate-specific antigen (PSA) level. Based on his age, family history, and prostate volume, his pretest probability for clinically significant prostate cancer is determined to be $20\%$ . To decide whether to perform an invasive transrectal ultrasound-guided prostate biopsy, the physician orders a novel urinary biomarker panel. The physician's clinical protocol dictates that a biopsy will only be deferred if a negative biomarker result reduces the patient's post-test probability of cancer to strictly less than $5\%$ .

The laboratory's traditional cutoff threshold for a 'positive' result on this biomarker panel yielded a sensitivity of $90\%$ and a specificity of $80\%$ . However, to reduce the hospital-wide burden of unnecessary biopsies in the general screening population, the laboratory's quality committee recently raised the cutoff threshold for a positive result. This adjustment increased the assay's specificity to $95\%$ but decreased its sensitivity to $76\%$ .

Based on the new laboratory threshold compared to the traditional one, which of the following best describes the mathematical impact on the interpretation of a negative test result for this patient?

The negative predictive value decreases, resulting in a post-test probability of disease greater than 5% and necessitating a biopsyAnswer
B
The negative predictive value increases due to the assay's higher specificity, providing greater reassurance to safely defer the biopsy
C
The likelihood ratio of a negative test decreases, enhancing the assay's ability to rule out cancer and allowing the biopsy to be deferred
D
The predictive values remain unchanged because they are intrinsic properties of the test, meaning the biopsy can still be deferred
E
The increased specificity proportionally decreases the false omission rate, rendering a negative result highly reliable for deferring the biopsy

Answer

The negative predictive value decreases, resulting in a post-test probability of disease greater than 5% and necessitating a biopsy.

Moving the cutoff threshold to increase a test's specificity requires stronger evidence (higher biomarker levels) for a positive result, which inevitably decreases the test's sensitivity. A decrease in sensitivity increases the false-negative rate. Mathematically, the negative likelihood ratio (

LR-

) is calculated as

(1 - \text{Sensitivity}) / \text{Specificity}

. The traditional

LR-

was

(1 - 0.90) / 0.80 = 0.125

. The new

LR-

(1 - 0.76) / 0.95 \approx 0.253

. Because the

LR-

has mathematically increased, a negative test result is now less capable of ruling out the disease. For this patient with a

20\%

pretest probability (pretest odds of

0.25

), the post-test odds of disease with a negative result rise from

0.031

0.063

. Converting back to probabilities, the post-test probability increases from

3.03\%

5.94\%

. Because this new probability exceeds the physician's strict

5\%

safety threshold, the biopsy can no longer be deferred. Consequently, the negative predictive value (NPV) of the test has decreased for this patient, demonstrating the clinical paradox that optimizing a test to reduce false positives can simultaneously force unnecessary invasive procedures on patients who test negative.

Step-by-Step Solution

Calculate the patient's pretest odds of having clinically significant prostate cancer.

Pretest odds =

0.20 / (1 - 0.20) = 0.20 / 0.80 = 0.25

Bayes' theorem requires pretest odds to calculate post-test odds using likelihood ratios.

Calculate the negative likelihood ratio (

LR-

) for both the traditional and new cutoff thresholds.

Traditional

LR-

(1 - 0.90) / 0.80 = 0.10 / 0.80 = 0.125

. New

LR-

(1 - 0.76) / 0.95 = 0.24 / 0.95 \approx 0.253

LR-

quantifies how much the odds of disease decrease when a diagnostic test result is negative.

Calculate the post-test odds for a negative result under both threshold conditions.

Traditional post-test odds =

0.25 \times 0.125 \approx 0.031

. New post-test odds =

0.25 \times 0.253 \approx 0.063

Post-test odds are the product of the pretest odds and the likelihood ratio.

Convert the post-test odds back to post-test probabilities and compare to the physician's

5\%

deferral threshold.

Traditional post-test probability =

0.031 / (1 + 0.031) \approx 3.03\%

(biopsy would be deferred). New post-test probability =

0.063 / (1 + 0.063) \approx 5.94\%

(biopsy is now required).

The clinical decision is based on an absolute probability threshold, which requires converting odds back into standard probability.

Key Concept

Impact of diagnostic cutoff threshold adjustments on likelihood ratios and post-test probabilities

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic