Overall, 108 consecutive patients (n = 49 females, 45.3%) with a mean age of 70.1 ± 14.3 years were included in the present study.

Table 1 displays the demographics of the investigated patients. Figure 1 shows exemplary cases of the patient samples. Overall, 43 patients were nodal positive (38.9%), whereas 65 patients were nodal negative (61.1%). N1 stage was found in 21 cases (19.4%), N2 in 22 cases (20.4%).

Representative cases of the patient sample. (a) 51-year old male patient, adenocarcinoma of the sigma, T3a, N0, M0, Node-RADS 1, short-axis-diameter 6 mm. (b) 78-year old female patient, adenocarcinoma of the distal colon transversum, T3b, N1b, M0, Node-RADS 5, short-axis-diameter 16 mm. (c) 50-year old female patient, adenocarcinoma of the colon descendens, T2, N1b, M0, Node-RADS 3, short-axis-diameter 9 mm

All tumours were adenocarcinomas. There were only 4 cases with a rare subtype, in two cases a mucinous adenocarcinoma and in two cases a signet-ring adenocarcinoma, respectively.

In 26 patients (24.1%) a metastasized stage was diagnosed with the following sites of distant metastasis, peritoneal in 14 cases (53.8%), liver in 13 cases (50.0%), lung in 4 cases (15.4%) and in case ovary (3.8%).

Among the metastasized patients 25 cases (96.1%) were nodal positive and one case was nodal negative.

When applying the commonly used threshold of 10 mm for short-axis-diameter to define malignant lymph nodes in clinical routine, 57/65 (87.7%) N- cases were correctly staged as N-, whereas 17/43 (39.5%) N + cases were correctly staged as N+. The resulting false positive rate is 8/65 (12.3%) and the false negative rate is 26/43 (60.4%).

When using a total Node-RADS-score of 4 or higher as threshold, 58/65 (89.2%) cases would have been correctly staged as N- and also 17/43 (39.5%) cases correctly as N+. The resulting false positive rate is 8/65 (12.3%) and the false negative rate is 26/43 (60.4%).

Using a fixed cut-off value for total Node-RADS-score of 3 and higher, an increase of correct staging for N + cases (24/43, 55.6%) can be accomplished, whereas over-staged cases inclined at almost the same rate (50/65, 77%). The resulting false positive rate is 15/65 (23.1%) and the false negative rate is 19/43 (44.1%).

The distribution of malignancy according to each Node-RADS score for both readers are shown in Table 2.

Node-RADS-scoring resulted for reader 1 in a total of n = 44 for Node-RADS 1 (40.7%), n = 25 for Node-RADS 2 (23.1%), n = 15 for Node-RADS 3 (13.9%), n = 10 for Node-RADS 4 (9.3%) and n = 14 (13%) for Node-RADS 5.

For reader 2, the results were n = 18 for Node-RADS 1 (36%), n = 12 for Node-RADS 2 (24%), n = 6 for Node-RADS 3 (12%), n = 10 for Node-RADS 4 (20%) and n = 4 (8%) for Node-RADS 5.

Inter-reader agreement was only fair for the Node-RADS scoring (k = 0.35, p < 0.001).

For reader 1, Node-RADS 1 had a malignancy rate of 29.5% and for reader 2 0%, Node-RADS 2 had a malignancy rate of 24% for reader 1 and of 25% for reader 2, Node-RADS 3 had a malignancy rate of 46.7% for reader 1 and 75% for reader 2. Node-RADS 4 yielded a malignancy rate of 50% for reader 1 and 35.3% for reader 2, for Node-RADS 5 reader 1 reached a malignancy rate of 85.7% and reader 2 71.4%.

In discrimination analysis, the total Node-RADS score showed statistically significant differences between N- and N + stage (for reader 1: mean 1.89 ± 1.09 score for N- versus 2.93 ± 1.62 score for N+, for reader 2: 1.33 ± 0.48 score for N- versus 3.65 ± 0.94 score for N+, p = 0.001, respectively).

ROC curve analysis for lymph node discrimination (N- versus N+) showed an area under the curve (AUC) of 0.68. A threshold value of 2 resulted in a sensitivity of 0.62 and a specificity of 0.71 (Fig. 2).

Results of the ROC curve analysis for discrimination of N- versus N + with total Node-RADS-Score, yielding an AUC of 0.68 [95 CI: 0.57–0.79]. A threshold value of 2 resulted in a sensitivity of 0.62 and a specificity of 0.71

Short axis diameter reached statistically significant difference between N- and N + stage (mean 6.8 ± 3.3 mm for N- versus 10.7 ± 7.3 mm for N+, p = 0.001).

Node-RADS subcategory size also reached statistically significant difference between N- and N + stage (mean 1.11 ± 0.31 score for N- versus 1.42 ± 0.54 score for N+, p < 0.001).

Also, for both readers, the Node-RADS-subcategories shape (p = 0.002) and border (p = 0.002) achieved statistically significance with higher subcategory-scores correlating with higher likelihood of positive N-stages. However, the subcategory texture did not reach statistical significance (p = 0.12).

ROC curve analysis for lymph node discrimination (N- versus N+) using the short axis diameter resulted in an area under the curve of 0.68. Using a threshold value of 6.5 mm, sensitivity reached 0.67 and specificity 0.52 (Fig. 3).

Results of the ROC curve analysis for discrimination of N- vs. N + with short axis diameter, the resulting area under the curve (AUC) is 0.68 [95% CI: 0.58–0.79]. Using a threshold value of 6.5 mm, sensitivity reached 0.67 and specificity 0.52

No significant difference was found between the AUC for lymph nodes discrimination using total Node-RADS score and the AUC for discrimination with short axis diameter (p = 0.85).

Inter-reader agreement was fair for the subcategory shape (k = 0.24, p = 0.001), moderate for the subcategories border (k = 0.44, p < 0.001) and texture (k = 0.52, p < 0.001), but reached only slight values for the added-up total configuration score, with a Cohen’s kappa of k = 0.18 (p < 0.001).