Efficacy and Safety of Gadobutrol 1.0 Molar (Gadovist) for Breast MRI

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the clinical investigators and the 3 blinded readers using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants. The difference was calculated as CMRM value minus UMRM value. For ease of expression, the following abbreviations will be used: Magnetic Resonance Mammography (MRM), Unenhanced MRM (UMRM), combined unenhanced and contrast (gadobutrol)-enhanced MRM (CMRM), X-ray mammography (XRM).

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the clinical investigators and the 3 blinded readers using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants.

A non-malignant breast was defined as false positive (FP), when the reader assessed at least one breast region as malignant. When all breast regions were assessed as non-malignant, the breast was defined as true negative (TN). Breast level specificity was first defined in participant as number of TN-breasts in participant divided by number of non-malignant breasts in participant. Subsequently the specificity percentage was calculated based on the mean of the specificities across all participants who contributed with at least one non-malignant breast.

A malignant breast was defined as false positive (FP), when the reader using the respective imaging modality assessed more breast regions as malignant as were present according to SoT. Otherwise the breast was assessed as true negative (TN). Specificity was then defined as TN/(TN+FP).

Index cancer is defined as the cancer confirmed by histology prior to inclusion which made the participants eligible for the study. The difference in percentage of participants was calculated as CMRM value minus UMRM value, CMRM value minus XRM value, CMRM value minus CMRM+XRM value respectively.

Additional cancer was defined as cancer which was present according to SoT, but which was not defined as index cancer, i.e. was not known when the participant was enrolled into the study. The difference in percentage of participants was calculated as CMRM value minus UMRM value, CMRM value minus XRM value, CMRM value minus CMRM+XRM value respectively.

A non-malignant breast was defined as FP when the reader assessed at least one breast region as malignant. A malignant breast was defined as FP, when the reader using the respective imaging modality assessed more breast regions as malignant as were present according to SoT. Otherwise the breast was assessed as TN. Specificity was then defined as (N-FP)/N, where N was total number of breasts.

For each region the reader chose the category which best described the extent of malignant disease, i.e. no, unifocal, or multifocal malignant breast disease. The proportion of correct matches of each defined image set to the SoT for the extent of malignant breast disease was referred to as the categorical accuracy. The majority read value for the 3 blinded readers was determined at the disease state level (no disease, unifocal, multifocal). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

For each region the reader chose the category which best described the extent of malignant disease, i.e. no, unifocal, or multifocal malignant breast disease. The proportion of correct matches of each defined image set to the SoT for the extent of malignant breast disease was referred to as the categorical accuracy. The majority read value for the 3 blinded readers was determined at the disease state level (no disease, unifocal, multifocal). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the reader using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants. The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for sensitivity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the reader using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants. The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for sensitivity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the reader using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants. The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for sensitivity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

A malignant breast was defined as FP, when the reader using the respective imaging modality assessed more breast regions as malignant as were present according to SoT. Otherwise the breast was assessed as TN. Specificity was then defined as TN/(TN+FP). The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for specificity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

A malignant breast was defined as FP, when the reader using the respective imaging modality assessed more breast regions as malignant as were present according to SoT. Otherwise the breast was assessed as TN. Specificity was then defined as TN/(TN+FP). The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for specificity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

A malignant breast was defined as FP, when the reader using the respective imaging modality assessed more breast regions as malignant as were present according to SoT. Otherwise the breast was assessed as TN. Specificity was then defined as TN/(TN+FP). The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for specificity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

For a single participant the sensitivity was defined as the proportion of malignant breast regions that were recognized by the reader using the respective imaging modality as malignant. Subsequently the sensitivity percentage was calculated based on the mean of the sensitivities across all participants. The majority read value for the 3 blinded readers was determined at the disease state level (evaluable regions for sensitivity). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave a different categorical determination, the majority response was the most severe disease category given by any of the 3 readers. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

The disease state "bilateral malignant disease" was derived from the assessment of the different regions for each breast (right and left) for investigators for each imaging modality (UMRM, CMRM, XRM, UMRM+XRM, and CMRM+XRM) based on the following rule: If the participant had at least one breast with no malignant region, the assessment of bilateral malignant disease was categorized as "No". If the participant had at least one malignant lesion in both breasts, the assessment of bilateral malignant disease was categorized as "Yes". The proportion of correct matches of each different image set to the SoT for the existence of bilateral malignant disease was derived. The analysis was based on the difference in accuracy for the evaluation of bilateral malignant disease for the following image comparisons on a participant level. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

The 3 blinded readers each recorded his/her confidence in diagnosis for each breast region based on a 4-point scale (1=not confident, 2=somewhat confident, 3=confident, and 4=very confident). The majority read value for the 3 readers was determined at the disease state level (no disease, unifocal, multifocal). If 2 of 3 or all 3 readers gave the same categorical determination of malignant disease for a breast region, the majority reader response was that category. If all 3 readers gave different categorical determination, the majority reader response was the most severe disease category given by any of the 3 readers, i.e. multifocal. For each participant, the mean of the confidence responses for the diagnosed breast regions was calculated, and rounded to the nearest 0.5. value respectively. The difference was calculated as CMRM value minus UMRM value, CMRM+XRM value minus UMRM+XRM value, CMRM+XRM value minus XRM value respectively.

Inter-reader agreement was assessed by considering each breast region to have 2 possibilities (malignant disease / no malignant disease) for an assessment by the 2 image sets (UMRM and CMRM). Kappa value varies from 0 (no agreement) to 1 (perfect agreement).

Intra-reader variability was assessed using a kappa on the match to SoT for the different regions within each participant (match, no match SoT). For each of the 3 readers, intra-reader agreement was assessed by considering each breast region to have 2 possibilities for an assessment by CMRM: matched SoT or did not match SoT. Kappa value varies from 0 (no agreement) to 1 (perfect agreement).

Intra-reader variability was assessed using a kappa on the match to SoT for the different regions within each participant (match, no match SoT). For each of the 3 readers, intra-reader agreement was assessed by considering each breast region to have 2 possibilities for an assessment by CMRM: matched SoT or did not match SoT. Kappa value varies from 0 (no agreement) to 1 (perfect agreement).

Intra-reader variability was assessed using a kappa on the match to SoT for the different regions within each participant (match, no match SoT). For each of the 3 readers, intra-reader agreement was assessed by considering each breast region to have 2 possibilities for an assessment by CMRM: matched SoT or did not match SoT. Kappa value varies from 0 (no agreement) to 1 (perfect agreement).

Systolic and diastolic blood pressure were measured in a supine position. Blood pressure was not to be measured on the arm used for the injection.

Heart rate was measured in a supine position.

Number of participants with at least one occurrence of changing from low or normal at baseline to high at follow-up.