Outcome
| Type |
Measure |
Description |
Time frame |
Safety issue |
| Primary |
Specificity of rapid on site evaluation (ROSE) of touch preparations (preps) for predicting malignancy |
Descriptive statistics (e.g., frequencies, ranges, means, medians, proportions, and standard deviations [SDs]) along with 95% confidence intervals (CIs) for the means or proportions will be computed for the measures of interest. Specificity will be defined as true negative (TN) divided by (TN + false positive [FP]). High probability of malignancy is defined as adequate tissue with tumor present. Low probability of malignancy defined is adequate tissue with no tumor present. Indeterminate probability of malignancy is defined as presence of atypical cells but inadequate for a definitive diagnosis on ROSE. Will collapse the indeterminate probability of malignancy (atypical cells category) found on touch preps into low probability for malignancy bin and indeterminate probability of malignancy on visual assessment into low probability for malignancy. |
Up to 1 year |
|
| Secondary |
Sensitivity of ROSE of preps for predicting malignancy |
Descriptive statistics (e.g., frequencies, ranges, means, medians, proportions, and SDs) along with 95% CIs for the means or proportions will be computed for the measures of interest. Sensitivity will be defined as true positive (TP) divided by (TP + false negative [FN]). High probability of malignancy is defined as adequate tissue with tumor present. Low probability of malignancy defined is adequate tissue with no tumor present. Indeterminate probability of malignancy is defined as presence of atypical cells but inadequate for a definitive diagnosis on ROSE. Will collapse the indeterminate probability of malignancy (atypical cells category) found on touch preps into low probability for malignancy bin and indeterminate probability of malignancy on visual assessment into low probability for malignancy. |
Up to 1 year |
|
| Secondary |
Specificity of visual assessment of pleura for predicting malignancy |
Descriptive statistics (e.g., frequencies, ranges, means, medians, proportions, and SDs) along with 95% CIs for the means or proportions will be computed for the measures of interest. Specificity will be defined as TN divided by (TN + FP). High probability of malignancy is defined as presence of abnormalities, such as studding, presence of parietal or visceral masses or nodules, or abnormal tissue deposits consistent with malignancy. Low probability is defined as absence of abnormalities or presence of purulent, fibrino-purulent pleural fluid or parietal and/or visceral pleura inflammation or thickening. Indeterminate probability is defined as findings which cannot include or exclude malignancy with certainty, such as inflammation of the pleura or adhesions, but that cannot be classified by the interventionist in any of the above categories. Will collapse the indeterminate probability of malignancy on visual assessment into low probability for malignancy. |
Up to 1 year |
|
| Secondary |
Sensitivity of visual assessment of pleura for predicting malignancy |
Descriptive statistics (e.g., frequencies, ranges, means, medians, proportions, and SDs) along with 95% CIs for the means or proportions will be computed for the measures of interest. Sensitivity will be defined as TP divided by (TP + FN). High probability of malignancy is defined as presence of abnormalities, such as studding, presence of parietal or visceral masses or nodules, or abnormal tissue deposits consistent with malignancy. Low probability is defined as absence of abnormalities or presence of purulent, fibrino-purulent pleural fluid or parietal and/or visceral pleura inflammation or thickening. Indeterminate probability is defined as findings which cannot include or exclude malignancy with certainty, such as inflammation of the pleura or adhesions, but that cannot be classified by the interventionist in any of the above categories. Will collapse the indeterminate probability of malignancy on visual assessment into low probability for malignancy. |
Up to 1 year |
|
| Secondary |
Positive predictive value (PPV) for all patients |
PPV will be defined as TP divided by (TP + FP). Pre-test odds will be calculated using the formula: pre-test probability divided by (1 - pre-test probability). Post-test odds will be calculated using the formula: pre-test odds times likelihood ratio, and post-test probability was calculated using the formula: post-test odds divided by (1 + post-test odds). Likelihood ratios (LRs) will also be calculated by dividing the probability of a result in patients with the disease by the probability of the same result in patients without the disease. |
Up to 1 year |
|
| Secondary |
Negative predictive value (NPV) for all patients |
NPV will be defined as TN divided by (TN + FN). Pre-test odds will be calculated using the formula: pre-test probability divided by (1 - pre-test probability). Post-test odds will be calculated using the formula: pre-test odds times likelihood ratio, and post-test probability was calculated using the formula: post-test odds divided by (1 + post-test odds). LRs will also be calculated by dividing the probability of a result in patients with the disease by the probability of the same result in patients without the disease. |
Up to 1 year |
|
| Secondary |
Specificity of ROSE on touch preps between centers |
Will be compared between centers using Chi-squared and Fisher exact tests. |
Up to 1 year |
|
| Secondary |
Sensitivity of ROSE on touch preps between centers |
Will be compared between centers using Chi-squared and Fisher exact tests. |
Up to 1 year |
|
