Formalin Fixation as Tissue Preprocessing for Multimodal Optical Spectroscopy Using the Example of Human Brain Tumour Cross Sections
Table 2
Overview of model quality parameters: ccuracy, sensitivity, specificity, false positive rate, and precision are calculated for each method and model.
Accuracy/%
Sensitivity/%
Specificity/%
False positive rate/%
Precision/%
Sample set I
UV absorption
Nonfixed
82
82
94
6
89
Fixed
95
95
98
2
96
Combined
91
91
97
3
91
Elastic light scattering
Nonfixed
88
88
96
4
88
Fixed
98
98
99
1
98
Combined
92
92
97
3
92
FTIR absorption
Nonfixed
93
93
98
2
94
Fixed
97
97
99
1
97
Combined
92
92
97
3
92
Sampel set II
UV absorption
Nonfixed
85
85
95
5
88
Fixed
75
75
92
8
79
Combined
84
84
94
5
84
Elastic light scattering
Nonfixed
72
72
91
9
74
Fixed
98
98
99
1
98
Combined
79
79
93
7
82
FTIR absorption
Nonfixed
87
87
96
4
87
Fixed
93
93
98
2
94
Combined
87
87
96
4
90
In general, the models of formalin-fixed cross sections are better than the nonfixed models. But, in combination, they are still working in one model independent from the preparation procedure.
