Regression Analysis for the Anomalies!

I have the ice drift anomalies(pixel value for one particular year- Avg value) data for the 36 winters; size 361*361*36

• Land areas have zero ice drift
• In earlier years some places has ice, so has the ice drift values but are ice free in the recent years

In my trend analysis it shows the negative trend for those areas, however i don't want to take those pixels in my analysis where there is no data for all the years. How to omit that.

%%Regression analysis for ice drift anomalies

   for i=1:361,
        for j=1:361,
           if (sum(isnan(anom_wint_r(i,j,:))> 0));
               idanom_trend_per_winter(i,j) = NaN;

            else

                stats = regstats(squeeze(anom_wint_r(i,j,:)),years','linear',{'beta' 'rsquare' 'fstat'});
                idanom_trend_per_winter(i,j) = stats.beta(2);
            end
        end
    end