There is no actual row object in Cassandra - rows don't exist themselves, they are just containers for columns. When querying, you can't tell the difference between a row with deleted columns and a row that never existed. So remove('row key') is doing exactly what you need to delete all columns in a row.
However, during range slice queries, you might see the row tombstone object - see http://wiki.apache.org/cassandra/FAQ#range_ghosts. This is a temporary object though and must not be used by applications to indicate anything.