new.profile: Create a new ultrasound profile

In Morphonodepredictivemodel/morphonode: The Morphonode Predictive Model for ultrasound signatures detection and malignancy prediction in vulvar cancer

Create a new ultrasound profile

Description

Creates a new ultrasound profile, used as input for the us.predict function.

Usage

new.profile(u = NULL, ...)

Arguments

u

A vector of length 14, corresponding to the input ultrasound profile. The order of the elements in u must be: short axis [numeric], cortical thickness [numeric], nodal core sign (hilum) [dichotomous], perinodal hyperechogenic ring [dichotomous], cortical interruption [dichotomous], echogenicity [dichotomous], focal intranodal deposit [categorical], vascular flow localization [categorical], cortical thickening [categorical], vascular flow architecture pattern [categorical], cortical-medullar interface distortion [categorical], shape [categorical], grouping [categorical], color score [ordinal]. Value -1 can be entered for missing values. If u = NULL, a guided interactive prompt will be launched. In the interactive mode, typing return without entering any value will introduce a missing value (this is not possible for short axis and cortical thickness, since they are necessary for a reliable prediction).

...

Currently ignored.

Value

A list of 2 objects:

1. "ultrasound", ultrasound features vector;

2. "missing", missing values vector (empty, if no missing values are found).

Author(s)

Fernando Palluzzi [e-mail]

See Also

See show.mpmenv for ultrasound variable description. See also us.predict to launch the full Morphonode Predictive Model suite.

Examples

### Profile with missing data

u <- new.profile(c(10.0, 6.3, 1, 0, 0, 0, -1, 1, 2, 2, 3, -1, -1, -1))
print(u)

### New profile from simulated data

# High metastatic risk profile
u.hmr <- new.profile(us.simulate(signature = "HMR"))
print(u.hmr)

# Low metastatic risk profile
u.lmr <- new.profile(us.simulate(signature = "LMR"))
print(u.lmr)

# Malignant profile
u1 <- new.profile(us.simulate(y = 1))
print(u1)

# Non-malignant profile
u0 <- new.profile(us.simulate(y = 0))
print(u0)

Morphonodepredictivemodel/morphonode documentation built on Feb. 15, 2023, 4:51 a.m.